Proles and training needs of fact-checkers in the Iberian Peninsula:
impact of AI
Perles y necesidades formativas de los fact-checkers de la Península Ibérica:
impacto de la IA
doxa.comunicación |
nº 40, pp. 465-491
| 465
January-June of 2025
ISSN: 1696-019X / e-ISSN: 2386-3978
How to cite this article:
Cantón-Correa, J.; Montoro-Montarroso, A.; Gómez-Romero, J. and Molina-Solana, M. (2025). Proles and training needs of fact-checkers in the
Iberian Peninsula: impact of AI.
Doxa Comunicación
, 40, pp. 465-491.
https://doi.org/10.31921/doxacom.n40a2725
Javier Cantón-Correa.
Associate researcher in the Department of Computer Science and Articial Intelligence at the
UGR and is a lecturer at UNIR, UOC and UI1. He holds a PhD in Social Sciences, a Master’s degree in Articial Intelligence
and Social Problems and a Bachelor’s degree in Sociology and Audiovisual Communication. He is an expert in social
networks, data analysis and visualisation. He has been a journalist at
RTVE.es
. He currently works on disinformation
and AI. Accredited as a Professor and Professor of Private University. One sexennium of recognised research. He is
currently a member of the Department of Computer Science and Articial Intelligence, and an associate researcher at
the Centre for Research in Information and Communication Technologies (CITIC-UGR).
University of Granada, Spain / International University of La Rioja, Spain
[email protected]
/
[email protected]
ORCID:
0000-0002-8466-1679
Andrés Montoro-Montarroso.
Researcher and lecturer in the Department of Information Technologies and Systems
at the University of Castilla-La Mancha (Spain) attached to the ITSI - Institute of Information Technologies and
Systems. He works on new methods and techniques for Computing with Words and Approximate Reasoning in Big Data
environments in SAFER and APIA projects. Previously, he was a researcher in the Department of Computer Science and
Articial Intelligence at the University of Granada (Spain), working on dierent projects of Articial Intelligence applied
to the prevention and detection of disinformation (IBERIFIER, XAI-disinfodemics, and U-MIND).
University of Castilla-La Mancha, Spain
[email protected]
ORCID:
0000-0003-1893-3346
Juan Gómez-Romero.
Professor at the Department of Computer Science and Articial Intelligence at the University of
Granada, Spain. Previously, he was a Lecturer and Researcher at Universidad Carlos III de Madrid, Spain. His research
focuses on applied articial intelligence, with applications in various elds such as disinformation characterization
and intelligent cyberphysical systems. Professor Gómez-Romero has contributed to numerous academic publications
and is actively involved in projects related to the development of AI-based solutions for media content analysis and
data extraction, particularly XAI-DISINFODEMICS and IBERIFIER Plus.
He is currently a member of the Department of
Computer Science and Articial Intelligence, and an associate researcher at the Centre for Research in Information and
Communication Technologies (CITIC-UGR).
University of Granada, Spain
[email protected]
ORCID:
0000-0003-0439-3692
is content is published under Creative Commons Attribution Non-Commercial License. International License CC BY-NC 4.0
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
466 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Recibido: 01/07/2024 - Aceptado: 29/09/2024 - En edición: 22/10/2024 - Publicado: 01/01/2025
Resumen:
El presente artículo analiza las necesidades y desafíos de los vericadores
de hechos en la Península Ibérica, en el marco del proyecto IBERIFIER. Este
estudio se centra en la identicación y evaluación de herramientas tecno
-
lógicas basadas en inteligencia articial (IA) utilizadas en la lucha contra
la desinformación en España y Portugal. Se destacan los perles de los ve
-
ricadores, las herramientas y ujos de trabajo actuales, así como las ne
-
cesidades de formación especícas. Se incluyen observaciones preliminares
de la dinámica laboral en cuatro redacciones de vericación en España,
subrayando las diferencias organizativas y metodológicas entre las entida
-
des. Además, se abordan las limitaciones y oportunidades que presenta la
implementación de soluciones de IA en este ámbito. Los resultados mues
-
tran que, aunque existen diversas herramientas tecnológicas, la eciencia
y efectividad de estas dependen en gran medida de la integración del factor
humano en el proceso de vericación. Finalmente, se proponen directrices
y mejores prácticas para mejorar la vericación de hechos mediante el uso
de tecnologías avanzadas, subrayando la importancia de la formación
continua y la colaboración entre diferentes actores para enfrentar la desin
-
formación de manera efectiva.
Palabras clave:
Inteligencia articial; vericación de hechos; desinformación; medios
digitales; Península Ibérica; España; Portugal.
Received: 01/07/2024 - Accepted: 29/09/2024 - Early access: 22/10/2024 - Published: 01/01/2025
Abstract:
is article analyses the needs and challenges of fact-checkers in the
Iberian Peninsula, in the framework of the IBERIFIER project. is
study focuses on the identication and evaluation of technological
tools based on articial intelligence (AI) used in the ght against
disinformation in Spain and Portugal. It highlights verier proles,
current tools and workows, as well as specic training needs.
Preliminary observations of the work dynamics in four verication
newsrooms in Spain are included, highlighting organizational
and methodological dierences between entities. In addition, the
limitations and opportunities of implementing AI solutions in this eld
are addressed. e results show that, although various technological
tools are available, their eciency and eectiveness depend mostly
on the integration of the human factor in the verication process.
Finally, guidelines and best practices are proposed to improve fact-
checking using advanced technologies, underlining the importance
of continuous training and collaboration between dierent actors to
eectively address disinformation.
Keywords:
Articial intelligence; fact-checking; disinformation; digital media;
Iberian Peninsula; Spain; Portugal.
Miguel Molina-Solana
. Associate Professor with the Department of Computer Science and AI, Universidad de Granada.
He was also a Marie Curie Research Fellow at the Data Science Institute, Imperial College London. His research
experience comprises work in Machine Learning and Knowledge representation and applied Articial Intelligence to
diverse domains such as Energy and Disinformation. He is currently a member of the Department of Computer Science
and Articial Intelligence, and an associate researcher at the Centre for Research in Information and Communication
Technologies (CITIC-UGR).
University of Granada, Spain
[email protected]
ORCID: 0000-0001-5688-2039
1. Introduction
e digital transformation of the media has generated new professional proles and competences in the eld of journalism,
especially in fact-checking. Fact-checking journalists have become key players in the ght against disinformation, requiring
advanced skills in the use of articial intelligence (AI) technologies and a deep understanding of the dynamics of digital
information. e increasing complexity of disinformation in the digital age demands that these professionals not only master
technological tools, but also understand user behaviour and the psychology behind the spread of false information (Agarwal
et al., 2023; Santos, 2023; Slapakova, 2021).
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 467
is study poses the following research questions to guide the analysis:
1. What are the main training needs of fact-checkers in the Iberian Peninsula in relation to the use of articial intelligence
tools?
2. Which technological tools are used by fact-checkers in Spain and Portugal, and how do they inuence their verication
practices?
3. What are the main limitations and opportunities in the implementation of AI technologies in fact-checking in the Iberian
region?
e specic objectives of this study can be found below:
–
To identify the AI-based technological tools used by fact-checking organisations in Spain and Portugal.
–
To assess the training needs of fact-checkers in the Iberian Peninsula.
–
To explore the role of AI in improving fact-checking practices.
–
To develop guidelines and best practices for the use of AI technologies in fact-checking.
–
To contribute to the strategic analysis of the impact of disinformation.
is article is structured in several sections that comprehensively address the issue of fact-checking in the Iberian Peninsula,
with a focus on the use of technological tools and articial intelligence. After an introduction that has contextualised the
importance of fact-checking in the current environment of disinformation, the article presents a theoretical framework that
reviews the existing literature on disinformation and verication tools. It then details the methodology employed in the
study, which combines qualitative and quantitative techniques to collect and analyse data. e results of this research are
then presented, followed by a discussion that interprets the ndings and suggests practical implications. Finally, conclusions,
limitations of the study and proposals for future research and improvements in the eld of fact-checking are presented.
2. eoretical framework
Disinformation, dened as the deliberate dissemination of false information for the purpose of misleading, has increased
its presence in the global media ecosystem, signicantly aecting public trust in the media and the perception of reality
(Fernández-Barrero et al., 2024). In the Iberian Peninsula, disinformation has had a notable impact in various elds, from
politics to public health, exacerbated by the COVID-19 pandemic. Disinformation not only undermines the credibility of
institutions, but also alters public perception and can inuence critical decisions. e proliferation of social media and digital
platforms has facilitated the rapid spread of false content, making it dicult for traditional media and fact-checkers to counter
these misleading narratives (Rodríguez Pérez, 2020). In Spain and Portugal, this situation has generated an urgent need to
develop eective fact-checking strategies.
On the other hand, technological tools play a crucial role in the ght against disinformation. ese tools include articial
intelligence (AI) algorithms for the automatic detection of fake news, collaborative verication platforms and OSINT or open-
source intelligence techniques (Esteban-Navarro et al., 2021). AI has been highlighted for its capacity, not only for journalism
468 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
e literature review shows that disinformation is a signicant challenge in the Iberian Peninsula, which requires the use of
advanced technological tools and close collaboration between various actors to be eectively combated. Continuous training
and media literacy are essential to improve the capacities of fact-checkers and society’s resilience in the face of disinformation.
While there are outstanding initiatives in the Iberian region, such as
Maldita.es
, Newtral or Polígrafo, it is crucial to consider
the global fact-checking landscape. In the United States, organisations such as
FactCheck.org
and PolitiFact have developed
advanced methodologies for fact-checking, integrating technological tools and citizen collaboration. In Latin America,
Chequeado has implemented innovative collaborative fact-checking strategies. ese international approaches reveal
common challenges, such as the need to develop specialised tools for emerging platforms and the management of large
volumes of data, as well as contextual dierences, especially in the use and acceptance of articial intelligence technologies in
dierent cultural and regulatory environments (Das et al., 2023). Comparing these initiatives allows identifying good practices
and adapting eective strategies in the Iberian context.
is comparative perspective also shows that, although there are universal challenges in the ght against disinformation,
such as the implementation of AI and the need for inter-institutional collaboration, signicant dierences are also identied
in each context. In the Iberian Peninsula, barriers to the adoption of these technologies include limitations in resources and
specialised training (Herrero & Herrera-Damas, 2021; Santos, 2023). On the other hand, in regions such as the United States
and Latin America, the use of automated tools is more established, but they face challenges related to public distrust and the
ethical implications of using AI in fact-checking (Zhou & Zafarani, 2021). Comparing these contexts would thus allow for an
assessment of whether the obstacles faced by fact-checkers in the Peninsula are specic to the region or reect structural
problems present in other parts of the world. Such an analysis would also suggest the need to adapt international best practices
to the cultural and operational particularities of the Iberian environment.
3. Methodology
To develop this work, a triangulated combination of qualitative and quantitative methods was used to collect and analyse
data (Bericat, 1998; Flick, 2012, 2015). ese methods included participant observation, online surveys and semi-structured
interviews, which allowed for a comprehensive and detailed insight into the needs and challenges of fact-checkers in the
Iberian Peninsula.
3.1. Participant observation
Participant observation was used in an exploratory way in this study, with the aim of gaining a preliminary insight into the
working environment of fact-checkers in the Iberian Peninsula. Although this qualitative technique allows researchers to
immerse themselves in the environment of the community under study and gain an in-depth understanding of its practices
and dynamics, its proper implementation requires time and successive visits to comprehensively capture the routines and
variations of the context under investigation (Angrosino, 2012). In this case, lack of time and resources prevented an adequate
number of visits, limiting the ability to obtain a comprehensive and detailed view of the daily practices of fact-checkers.
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 469
It would also be useful to have the active engagement of fact-checking organisations from the planning phase of the study,
ensuring their collaboration to motivate the participation of their journalists. Finally, oering incentives that recognise the
time and eort put in by participants could increase the response rate and, consequently, the validity of the ndings.
e survey design was designed to identify the technological tools used, training needs and specic challenges faced by
verication professionals in their daily work. e survey consisted of 30 questions, divided into three main sections:
–
Professional prole and experience: Questions on the fact-checkers’ work history were included, such as years of experience
in the eld, type of training received and perception of their professional competence.
–
Tools and workows: is section explored the technological tools that fact-checkers commonly use, as well as the main cha
-
llenges they face in their use. Questions were asked, for example, about the use of specic tools for searching and archiving
information and about the platforms they consider most relevant for their work.
–
Training needs and challenges: Data were collected on fact-checkers’ perceptions of their training needs and the diculties
they encounter in their work. It included open questions on training needs and challenges in the use of technological tools
for fact-checking.
Although the analysis of the data obtained made it possible to identify patterns and trends among the fact-checkers who
participated in the survey, the low response rate (10%) signicantly limits the ability to generalise these results to the entire
population of verication journalists in the Iberian Peninsula and Ibero-America. is non-response bias suggests that the
ndings should be interpreted with caution, as respondents may not adequately reect the diversity and needs of the entire
population. Still, the results provide useful preliminary insights that contribute to understanding the dynamics and challenges
of the eld but cannot be considered representative or conclusive.
To address these limitations in future research, it is recommended to implement strategies to increase the response rate,
such as the use of personalised reminders and the active involvement of fact-checkers. In addition, comparative analyses
between the characteristics of respondents and non-respondents would be benecial to assess possible biases and improve
data interpretation.
e surveys included closed and open-ended questions to collect both quantitative and qualitative data, allowing for an in-
depth exploration of the technological tools used, workows, training needs and specic challenges faced by fact-checkers in
their daily work. Although this type of survey was selected for its eciency in obtaining information from many participants
(D’Ancona, 1996), the low response rate obtained (10%) limits the ability to generalise the results to the entire population of
fact-checkers in the region. e ndings provide a preliminary insight into the trends and needs of the eld, but they should
be interpreted with caution, as they cannot be assured to represent the full reality of the eld.
Despite these limitations, the information collected provides a valuable starting point for future research and the development
of training strategies. It is important to note that the data obtained is available upon request and under conditions that
ensure its responsible and justied use for research purposes. In future studies, it is recommended to consider additional
methodologies and direct collaboration with the verication bodies to increase participation and obtain more representative
and robust results.
470 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
4. Results
4.1. Preliminary results of participant observation in verication editorial oces
e participant observation carried out in four verication oces in Spain provided a preliminary insight into the dynamics
and methodologies employed in each organisation. Although it was an exploratory analysis with time and scope limitations,
the results allow us to identify relevant patterns and dierences between the organisations studied.
EFE Verica is characterised by its public service-oriented approach and editorial independence. Despite having a solid
organisational structure, it faces challenges related to resource management and the need to respond quickly to issues of high
virality, which may limit the depth of the verications carried out.
Maldita.es
stands out for its ability to innovate in the use of technological tools and its focus on collaboration with citizens. e
entity has developed eective strategies to interact with its audience through platforms such as WhatsApp, which allows it to
identify and respond to rumours and disinformation in real time. However, prioritising issues that are not necessarily viral but
socially relevant remains a constant challenge.
Newtral presents a team highly specialised in the use of data and emerging technologies. Its organisational structure favours
collaboration between its research and production teams, facilitating the integration of advanced methodologies into its
workows. However, the adoption of new tools and the management of large volumes of information in real time continue to
be areas for improvement.
VericaRTVE, being part of a public service entity with access to mass dissemination resources and platforms, has a signicant
capacity to address relevant and current issues. However, its hierarchical structure and the need to coordinate with other
internal areas slow down the speed of response and publication of verications, which may impact the eectiveness of its
work.
ese observations reect the diversity of approaches and challenges in the fact-checking ecosystem in Spain, highlighting
the need for specic technological and organisational adaptations to optimise the response to disinformation. Although the
results obtained should be interpreted with caution due to their exploratory nature, they oer a starting point for future more
in-depth and systematic research in this eld.
4.2. Survey on the status and technology needs of information fact-checkers
e main results obtained from the surveys and interviews with fact-checkers in Spain and Portugal are presented. Firstly,
the data collected from the surveys are analysed. Subsequently, the ndings derived from the semi-structured interviews are
detailed.
e rst block of questions in the survey was aimed at nding out the prole of age, gender and education. e average age of
the fact-checkers who participated in the survey shows signicant dierences according to gender. Male fact-checkers have a
greater variability in age, ranging from 26 to 65 years, with a median age of 37.5 years. In contrast, the median age of women
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 471
In contrast, 43% of women have 2 years or less experience, compared to 24% of men. Many respondents (48%) consider
themselves to be professional fact-checkers, while 24% do not consider themselves to be professionals despite having sucient
experience, and 5% consider themselves to be beginners. is suggests a high level of condence in their skills and knowledge,
possibly due to the training and experience acquired.
On the tools and workows specic to journalist fact-checkers, the survey revealed that the most important social networks
for fact-checkers in their work (Figure 2) are X (formerly known as Twitter), WhatsApp and Instagram. X is the most used by all
respondents, followed by WhatsApp (76%) and Instagram (71%). Telegram and TikTok (66% each), and YouTube and Facebook
(62% each) are also mentioned. e least important are VK (14%) and Twitch (4.6%).
Figure 2. Digital platforms important for fact-checkers
(multiple answer)
Source: own elaboration
Fact-checkers face several challenges in detecting disinformation on dierent platforms. WhatsApp and Telegram present
diculties due to the private nature of communications, making it dicult to trace the source of disinformation. TikTok is
472 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
WhatsApp is the most used internal communication tool (71%), followed by email (52%). Other tools such as Slack, Signal,
Teams, Telegram and Google Meet are used less frequently. e preference for WhatsApp is due to its ease of use and
accessibility, while Signal stands out for its security and privacy. e adoption of tools such as Slack could be better integrated
into organisational communication.
Fact-checkers prefer specialised web-based information search and archiving tools (Figure 4), such as Wayback Machine and
Search by Image (67% each). TinEye and Google Maps/Earth are also popular (62%). Tools such as Archive.today, InVID and
CrowdTangle are used by 52% of fact-checkers, while TweetDeck is used by 48%. Tools related to articial intelligence and
other specic tools have signicantly lower usage, indicating a need for more training and resources in these areas.
Figure 4. Digital work tools used in the verication of information
(multiple answer)
Source: own elaboration
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 473
Figure 5. Training needs of Iberian fact-checkers
(multiple answer)
Source: own elaboration
As shown in Figure 5, a remarkable 76% of respondents showed interest in being trained in Articial Intelligence (AI)
techniques, highlighting their growing relevance in information verication. 66% are interested in OSINT (open-source
intelligence) techniques, and 62% want to go deeper into data and information analysis, using languages such as Python or R.
In addition, 57% want training in image analysis, and 48% in data visualisation. ey are also interested in specic fact-checker
techniques (43%) and in information search and retrieval, video analysis and social media (33%). Finally, 24% expressed
interest in improving their information dissemination skills, highlighting the need to manage social media platforms and
digital marketing tools to maximise the impact of veried information.
In terms of the use of training in specic tools, fact-checkers have identied several key needs in the eld of verication that
are not being addressed. ere is signicant demand for specialised tools for social media, image and video analysis, as well
as the use of articial intelligence (AI). e need for tools to monitor TikTok was also highlighted, given its exponential growth
and the video-centric nature of the platform. In addition, the importance of web scraping tools to extract data from various
474 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
a valuable tool, but it cannot replace human judgement and expertise’ (I4, p. 5), or put another way: ‘although AI can help in
detection, the nal decision on whether a content is disinformation or not still depends on human judgement’ (I7, p. 5).
4.3.2. What role can AI play in the fight against disinformation?
Disinformation is a growing phenomenon that threatens the integrity of information and public trust in news sources. In
this context, Articial Intelligence (AI) emerges as a powerful tool to combat the spread of fake news and ensure that
the information we consume is accurate and trustworthy. AI has a wide eld to explore in task automation, especially in
monitoring, where it can be more ecient than human supervision by crawling and analysing large amounts of data in real
time, identifying potential hoaxes or disinformation before they become widely disseminated: ‘AI allows us to monitor more
eciently, detecting hoaxes in real time’ (I1, p. 2).
In addition, AI can cluster verications, recognising recurring hoaxes and alerting fact-checkers about information already
identied as false, enabling a faster and more ecient response. For example, hoaxes updated with new information but
maintaining the same basic structure: ‘AI helps us identify hoaxes that are repeated or slightly modied’ (I1, p. 2). is
‘matching’ capability means that fact-checkers do not have to repeat the verication work every time they come across a
similar claim: ‘this tool will count our already published verications and tell us that we have already veried it and it is
false, and this politician has not been corrected or they keep repeating this myth, establishing the idea that it is true and you
should attack it harder’ (I3, p. 2). It can also help to detect and recognise disinformation strategies, identifying patterns in the
dissemination of disinformation, especially when certain data, such as unemployment, are manipulated by dierent political
actors: ‘to detect strategies of deception, [...] which allows us to cluster it so we don’t have to do the work of verication every
time we receive one of these messages’ (I3, p. 2). is ability to identify ‘clusters’ of disinformation is especially useful when
dierent actors repeat the same disinformation.
4.3.3. What are the challenges in implementing automation through AI?
e implementation of automation through AI in the eld of fact-checking and journalism presents several challenges. One
of the main ones is the tension between automation and the need for human intervention. While AI can handle repetitive,
high-volume tasks, human interpretation and judgement remain essential in many aspects of fact-checking, but one can try
to reinterpret this symbiotic relationship: ‘we have a lot of room to grow in the area of task automation, there is still too much
human presence, [which] is easily solvable with existing technologies that we have not developed in the right way’ (I1, p. 2).
e ability to monitor and detect hoaxes or fake news in real time is essential, and AI can help in this process, although there is
still room for improvement in terms of accuracy and scope: ‘we have a lot of growing to do, especially in the area of monitoring’
(I2, p. 2).
Moreover, dierentiating between recurrent and new deceptions is a challenge. AI can help identify patterns, but the nal
interpretation often requires a human touch: ‘it has to do with automated work, which [...] we do not attention to because
we have already debunked it and people don’t need to know that we have debunked it before’ (I3, p. 2). Despite advances in
AI, human supervision remains essential to ensure accuracy and quality of verication: ‘we will always need some human
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 475
In this scenario, articial intelligence (AI) emerges as a promising tool for both the creation and detection of disinformation.
AI enables the automation of complex tasks that previously required human intervention, facilitating the generation and
dissemination of false content, known as deepfakes or ultra-fakes, which include texts, images, audios and videos manipulated
with a high degree of realism (Bontridder & Poullet, 2021). However, the same technology that amplies disinformation can
also be harnessed to combat it. AI-based solutions oer the possibility of detecting false information more accurately and
eciently, overcoming the limitations of traditional manual methods (Montoro-Montarroso et al., 2023).
Fact-checkers in the Iberian Peninsula, within the IBERIFIER project, face specic challenges that require a combination of
advanced digital competences and a robust ethical approach. Recent studies have highlighted the importance of developing
technological skills and new professional competences in journalism to eectively address disinformation (López-García
et al., 2017; Martín-Antoranz et al., 2019). ese new professional proles must be able to use AI technologies to automate
and improve the accuracy of fact-checking, while maintaining critical and ethical judgement in the evaluation of information
(Cifuentes Arias & Sixto-García, 2022; Lazo et al., 2020).
e ght against disinformation has become crucially important in the current context, marked by the proliferation of
information on digital platforms and social networks. Disinformation, dened by the European Commission as ‘false,
inaccurate or misleading information designed, presented and promoted to cause public harm or economic gain’ (European
Commission, 2018), has proven to have severe consequences on public perception of critical issues, from politics to public
health. Notable examples include the rapid spread of misinformation during the COVID-19 pandemic, underlining the need
for reliable, fact-based information for informed decision-making and the maintenance of democracy.
In addition, the digitisation of journalism has changed the relationship between journalists and their audience, making the need
for transparency and trust in the information provided more evident (Arcila et al., 2020). In this context, verifying journalists
must play an active role not only in detecting disinformation, but also in educating the audience about the importance of
consuming veried and reliable information.
is work is also part of the IBERIFIER
1
project, simultaneously part of the European Digital Media Observatory (EDMO),
which aims to analyse the digital media ecosystem in the Iberian Peninsula and address the problem of disinformation through
research, fact-checking and the development of advanced technological tools, through the association of various entities in a
multidisciplinary consortium.
is article, which is the result of the work of this project, focuses on identifying and evaluating the AI-based technological tools
used in Spain and Portugal, as well as on highlighting the proles of fact-checkers and their training needs. rough this study,
we seek to provide a comprehensive understanding of the challenges and opportunities presented by the implementation of
AI solutions in the ght against disinformation, exploring how the integration of AI technologies can improve the practices of
fact-checkers in the Iberian Peninsula and address the current challenges of disinformation (Berganza et al., 2017; Sánchez-
García et al., 2023).
1
https://iberier.eu/
<p>[Enlace de URL / hc (has AS)]</p>
476 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
in general (Cantón-Correa et al., 2024), but also for analysing large volumes of data and detecting patterns indicative of
disinformation (Cantón-Correa, 2023), although its implementation poses ethical and technical challenges that require
human oversight to ensure accuracy and proper context of information (Montoro-Montarroso et al., 2023). In addition, OSINT
techniques allow fact-checkers to collect and analyse information from a variety of open sources, such as social networks and
public databases. ese techniques are essential for tracing the source of disinformation and understanding its spread, thus
improving the eectiveness of verication processes (Vizoso & Vázquez-Herrero, 2019).
In the Iberian Peninsula, several local initiatives have emerged to address disinformation. Organisations such as
Maldita.
es
, Newtral in Spain, or Polígrafo in Portugal, have stood out for their eorts in fact-checking and media education. ese
organisations use advanced technological tools and promote collaboration between journalists and the public to combat
disinformation (Lotero-Echeverri et al., 2018). Despite technological advances, fact-checkers face multiple challenges. e lack
of specialised tools for social network analysis and monitoring new platforms such as TikTok is a signicant limitation (Herrero
& Herrera-Damas, 2021). In addition, there is an urgent need for continuous training in the use of advanced technologies and
in understanding the dynamics of disinformation (Rodríguez Martínez et al., 2021).
e literature also highlights the importance of collaboration between dierent actors, including the media, fact-checking
organisations and the public, to create a united front against disinformation (Alfonso et al., 2021). is implies not only the
development of technological tools, but also the promotion of media literacy so that citizens can eectively identify and reject
disinformation (Ufarte-Ruiz et al., 2018).
In this paper, we will conceptualise disinformation as the deliberate dissemination of false information with the aim of
misleading or manipulating public opinion. A phenomenon that has been exacerbated using social networks and other digital
platforms, which allow for a rapid spread of false content (Echevarría, 2016). On the other hand, we will understand fact-
checking as the process of verifying the veracity of published information, especially in the media and social networks. is
process includes the identication of veriable facts, the search for evidence and the publication of the verication results
(Lotero-Echeverri et al., 2018). In this way, journalists specically dedicated to the task of fact-checking (which is an integral
part of journalism itself) are what we commonly known as fact-checkers.
e literature also highlights the importance of collaboration between dierent actors, including the media, fact-checking
organisations and the public, to create a united front against disinformation (Alfonso et al., 2021). is implies not only the
development of technological tools, but also the promotion of media literacy so that citizens can eectively identify and reject
disinformation (Ufarte-Ruiz et al., 2018).
In this paper, we will conceptualise disinformation as the deliberate dissemination of false information with the aim of
misleading or manipulating public opinion. A phenomenon that has been exacerbated using social networks and other digital
platforms, which allow for a rapid spread of false content (Echevarría, 2016). On the other hand, we will understand fact-
checking as the process of verifying the veracity of published information, especially in the media and social networks. is
process includes the identication of veriable facts, the search for evidence and the publication of the verication results
(Lotero-Echeverri et al., 2018). In this way, journalists specically dedicated to the task of fact-checking (which is an integral
part of journalism itself) are what we commonly known as fact-checkers.
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 477
Visits were made to four newsrooms (Maldita, Newtral, EFE Verica and VericaRTVE) in Spain, where interactions and
workows were observed. is exploration, carried out over the course of a week in October 2022, provided a valuable window
into the internal operations of these four leading newsrooms. During these visits, there was an opportunity to observe rst-
hand the dierent workows and realities of each entity, which gave an insight into the specic methodologies and work
dynamics that characterise each organisation. is rst-hand experience provided an in-depth and diverse perspective of the
information verication landscape in the current context.
However, it was not possible to access Portuguese fact-checkers’ editorial oces, which further limited the perspective of the
Iberian fact-checking ecosystem. ese circumstances, together with the exploratory nature of the technique used, restricted
the depth and scope of the analysis. erefore, the ndings should be interpreted as preliminary approximations and not
as conclusive results. ese initial ndings, although exploratory, provide an essential starting point for developing more
detailed and comprehensive fact-checking studies in the region.
Despite the exploratory nature of participant observation, the results obtained oer a rst insight into the specic dynamics
and practices of fact-checking newsrooms in Spain. In previous studies, this technique has proven useful to understand
everyday practices and discrepancies between what is reported and what happens in journalistic contexts (Emerson et al., 2011;
Hammersley & Atkinson, 2019; Quian, 2021; Spradley, 1980). Despite the limitations, the preliminary ndings are valuable to
better contextualise the results obtained with the other techniques used in this study. is exploratory approach also allows
for the identication of key areas for future research, where a more systematic and prolonged approach is recommended,
including successive visits and access to additional newsrooms, such as those in Portugal, which could not be studied on this
occasion.
3.2. Online surveys
To obtain a quantitative overview and to cover a wider sample, an online survey was designed for verication professionals
in dierent regions. e survey was conducted through the Google Forms tool and is available in the annex of the report. It
was sent by email to 205 journalists working in verication within specialised organisations in Ibero-America (a list of 217
verication journalists from Spain, Portugal and Latin America was located and compiled). A total of 21 responses were
obtained, which represents about 10% of the surveyed population.
is low response rate introduces a non-response bias that may aect the external validity of the results (Groves, 2006). is
bias is due to several factors, such as fact-checkers’ work overload and lack of time to participate in the research. e list of
journalists was compiled from data available on the websites of their respective organisations and from publicly accessible
information. e survey was sent out at dierent times or waves, trying to improve the response rate, but without signicant
success. is may be due to the workload associated with fact-checking, which requires constant attention to current events,
leaving little time to engage in external research. In addition, other factors such as the perceived relevance of the study or the
absence of incentives to respond may have inuenced low participation.
To mitigate this bias and improve representativeness in future research, it is recommended to implement additional strategies,
such as personalised reminders, complementary short interviews or the combination of qualitative and quantitative methods.
478 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
3.3. Semi-structured interviews
As a complement to the previous techniques, semi-structured interviews were conducted with the heads of dierent verication
entities, as well as with various journalists dedicated to research tasks within these entities. Semi-structured interviews are a
qualitative research method that combines elements of structured and unstructured interviews. ey allow the researcher
to address specic questions and, at the same time, oer sucient exibility for the interviewee to respond in detail and
share additional information that may arise during the conversation (Kvale, 2012). ey oer the following methodological
advantages:
–
Depth of information: Semi-structured interviews oer the opportunity to obtain more detailed and contextualised responses
than, for example, surveys. In addition, they allow interviewers to probe and probe more deeply into specic aspects.
–
Flexibility: In contrast to fully structured interviews, semi-structured interviews allow questions to be adapted and adjusted
according to the interviewee’s answers and the dynamics of the conversation.
–
Personal connection: ese interviews allow for a more direct and personal dialogue, which can lead to greater openness and
sincerity on the part of the interviewee.
–
Contextual adaptability: Depending on the prole of the interviewee, researchers can adapt the interview to maximise the
relevance and pertinence of the exchange.
–
Complementary to other methodologies: While participant observation and surveys provide general and specic insights,
semi-structured interviews allow for a deeper focus on individual perspectives.
e interviews were conducted between September and December 2022, with 10 key representatives from dierent verication
entities, namely Vericat, Maldita, EFE Verica, Newtral and VericaRTVE in Spain, and Lusa and Poligrafo in Portugal. Each
interview was designed considering the prole and responsibilities of the interviewees, thus ensuring that relevant and
detailed data was obtained from each entity. e combination of quantitative and qualitative techniques guarantees a holistic
and detailed view of the current situation of the fact-checkers.
3.4. Methodological combination
e combination of qualitative and quantitative methods was selected to ensure a comprehensive and nuanced understanding
of the challenges and needs of the fact-checkers. Participant observation provided, even with the limitations, a practical
and contextual insight into the workows and tools used, while online surveys allowed for ecient data collection from a
broad group of participants. Semi-structured interviews complemented these methods by oering additional depth and the
opportunity to explore emerging issues in greater detail.
ese combined methods ensured robust and comprehensive data collection, enabling researchers to develop
recommendations informed by a wide range of perspectives and experiences.
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 479
is 28, with the majority being under thirty. is suggests that many male journalists have specialised in fact-checking at some
point in their careers, while many women have entered journalism directly through fact-checking.
e educational level of the fact-checkers surveyed (Figure 1) is high: 95% have some kind of university degree, 57% have a
specialised Master’s degree and 33% have a university certicate. ere is a gender dierence: more men (33%) have a Master’s
degree than women (24%), and more women (24%) have a university degree than men (9.5%). Many fact-checkers (43%) have
between 1 and 2 years of experience, suggesting that information verication is a growing profession, driven by increasing
demand, training programmes and the availability of verication tools.
Figure 1. Highest educational attainment of fact-checkers, by gender
4.2
. Survey on the status and technology needs of information
fact-checkers
The main results obtained from the surveys and interviews with fact
-
checkers in Spain and
Portugal are presented.
Firstly
, the data collected from the surveys are analysed.
Subsequently, the findings derived from the semi
-
structured interviews are detailed.
The first block of questions in the survey was aimed at finding out the profile of age,
gender and education. The average age of the
fact-checkers
who participated in the survey
shows significant differences according to gender. Male
fact-checkers
have a greater
variability in age, ranging from 26 to 65 years, with a median age of 37.5 years. In contrast,
the median age of women is 28, with the majority being under thirty. This suggests that
many male journalists have specialised in fact
-
checking at some point in their careers,
while many women have entered journalism directly through fact
-checking.
The educational level of the
fact-checkers
surveyed (Figure 1) is high: 95% have some
kind of university degree, 57% have a specialised Master
’
s degree and 33% have a
university certificate. There is a gender difference: more men (33%) have a Master
’
s
degree than women (24%), and more women (24%) have a university degree than men
(9.5%).
Many
fact-checkers
(43%) have between 1 and 2 years of experience, suggesting
that information verification is a growing profession, driven by increasing demand,
training programmes and the availability of verification tools.
Figure
1.-
Highest educational attainment of
fact-checkers
, by gender
Source: own elaboration
By gender, it is observed that men tend to be more experienced than women, consistent with the initial age and gender
distribution. 24% of men have more than 3 years of experience, twice as many as the 9.5% of women with such experience.
480 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
also dicult to monitor due to its opacity and limitations in advanced search tools. Some fact-checkers do not have signicant
problems on more common platforms, but encounter diculties on less well-known digital media.
e main tasks of information fact-checkers include monitoring social networks and keeping up with current news (Figure
3). Approximately 91% conduct information and documentation searches, contact and interview experts, and write articles.
Around 81% participate in editorial meetings and use digital tools to search for evidence and publish articles. Only 5% engage
in video editing and training on disinformation, suggesting specialisation in these areas.
Figure 3. Work tasks carried out for the verication of information
(multiple answer)
Source: own elaboration
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 481
Fact-checkers rate the reliability of the tools they use (4.5 on average), followed by ease of use (4.3) and time eciency (4.2).
However, features such as privacy, learning curve and user interface receive lower scores. e lack of use of programming
languages in fact-checking work stands out, except for 5% who use HTML and CSS for web layout.
Fact-checkers face a variety of diculties in their work, according to responses to open-ended questions in the survey.
e main challenging tasks include verifying sources, searching for claims on podcasts and YouTube, searching for ocial
documents, and verifying videos and images, due to the time required. Managing and monitoring large volumes of data and
information also prove to be cumbersome tasks, essential for real-time verication. In addition, specic challenges such as
the automatic detection of claims in podcasts and the attractive presentation of information were mentioned, underlining the
complexity and diversity of tasks in fact-checking.
To investigate the training needs of the fact-checkers, questions were asked in the survey about the tools and programming
languages they use, as well as the diculties they face in their work. 76% of the participants have received training in fact-
checking and 100% expressed interest in further training in the future. Training is crucial to ensure accuracy and reliability
in journalism. e results show a diversity in the sources and modalities of training, with formal oine and online education
getting the same percentage (33%), and on-the-job training being the most popular (43%).
e last block of the survey contained questions about the training needs that the fact-checker journalists recognised. In this
respect, only 5% have received training on fact-checking as part of their university degree, and 9.5% have taken a specic
Master’s degree on the subject. is reects a lack of integration of fact-checking into traditional university curricula, although
its importance is recognised at postgraduate level. Furthermore, 24% of respondents have self-trained through online
resources and practical on-the-job training, while 19% have done so through oine resources such as books, highlighting the
importance of individual initiative and adaptability in a constantly changing eld.
482 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
online sources and easier access to ocial sources, preferably using AI, was mentioned. Finally, the need for an advanced
video editor with animation features to produce engaging and eective content was expressed.
In terms of emerging trends in disinformation that respondents would like to address, the need for continuous monitoring
of sources is highlighted, given that sources of disinformation are constantly changing and evolving. Also of concern is the
diversication of channels through which disinformation is distributed, requiring fact-checkers to navigate and monitor
multiple platforms. Another emerging issue is greenwashing, the practice of exaggerating the environmental credentials of an
organisation, product or service. In addition, there is signicant concern about articial intelligence, including the limitations
and future of AI, deep fakes and malicious use of AI, underscoring the need for a deep understanding and ability to counter
these technological challenges.
Finally, respondents identied several unmet technical needs in newsrooms. One participant mentioned the absence of
adequate fact-checker tools, indicating the need for more concrete, eective and ecient tools. e need for an accessible
and reliable database of experts that fact-checkers can quickly consult was also identied. In addition, the need for specialised
climate change editors and the importance of a robust and user-friendly video editor were mentioned. Finally, the need for
better nancial resources to support fact-checking work was highlighted, as this process can be resource intensive and without
adequate funding, newsrooms may not be able to invest in the tools, training and sta needed to do this work eectively.
4.3. Individual interviews with fact-checker journalists
4.3.1. What is the status of the use of AI in the fight against disinformation?
Articial intelligence (AI) has emerged as a promising tool in the ght against disinformation and the spread of fake news. e
interviewees, experts and heads of fact-checkers, oer a diverse but complementary perspective on the role of AI in this ght.
Overall, AI is recognised as having signicant potential to improve eciency and accuracy in detecting fake news, highlighting
its ability to process large amounts of data and detect patterns that are dicult for a human to identify. ‘AI has the potential
to be a powerful tool, but we are still in the early stages of its implementation’ (I1, p. 3). e importance of combining AI
with human judgement is stressed, as technology alone cannot fully understand the context and subtleties of language. Some
interviewees (I5, I6 and I7) mention the usefulness of AI-based tools for tracking and monitoring social networks and other
media, but also point out that these tools still have room for improvement.
However, concerns and reservations are also expressed: although AI can help in detection, the nal decision on whether a
piece of content is disinformation or not still depends on human judgement (I5, I7). ‘AI is a valuable tool, but it cannot replace
human judgement and expertise’ (I4, p. 5). Furthermore, several interviewees mention the ethical challenges and possible
unintended consequences of relying too much on AI in news fact-checking: ‘we need to be cautious and make sure that AI is
not misused’ (I3, p. 4). In conclusion, although AI is being used to some extent in the ght against disinformation, there is still
much room to expand its use and improve its eectiveness. AI is seen as a valuable tool that can complement and enhance
fact-checking eorts, but it is essential to approach it with caution and ensure its ethical and responsible use. Combining AI
with human judgement seems to be the most eective approach in the ght against disinformation, as it is mentioned: ‘AI is
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 483
supervision’, understanding verication ‘never as something that could be fully automated’ (I4, p. 2). Implementing AI
solutions in practice can be challenging, especially when it comes to new or untested tools: ‘[there are] tools that due to their
youth do not yet have [...] additional user involvement’ (E5, p. 2). In addition, advanced AI tools often have a cost associated
with them, which can be a barrier for organisations with limited resources: ‘many of these tools are paid tools’, for which
OSINT tools can be a way out, but not the only way out, using ‘the OSINT environment as an additional leg, but also entering
fully into the world of paid tools’ (I6, p. 4). ese challenges reect the considerations practitioners face when implementing
AI solutions in the eld of fact-checking and journalism. While AI oers signicant opportunities to improve and automate
processes, it also presents challenges that must be addressed to ensure its eectiveness and accuracy.
5. Discussion
e survey data reect an interesting trend in fact-checking in Spain. e predominant youthfulness of female fact-checkers
suggests a possible generational renewal in journalism, with women entering directly into specialised fact-checking roles.
e high educational level of the fact-checkers indicates that this profession requires a solid academic background, which
may contribute to the high level of condence they have in their skills. In addition, the fact that fact-checking is an emerging
profession, with many fact-checkers having little experience but already considering themselves professionals, suggests that
the demand for these skills is driving rapid professionalisation of the eld. is dynamic reects a response by the journalism
sector to growing concerns about disinformation and the critical need for fact-checking in the digital age.
Although participant observation was limited in time and scope, it provided relevant information about the internal dynamics
of the newsrooms observed. Dierences were observed in the organisation of teams and the use of technological tools among
the entities visited. ese ndings, although preliminary, complement the interpretation of the quantitative and qualitative
data and suggest the need to adapt the fact-checker strategies to the specic organisational characteristics of each entity.
On the other hand, fact-checkers in Spain and Portugal share similar objectives, although the media and socio-political
contexts present signicant dierences. In Spain, high political polarisation (Torcal, 2023) and the proliferation of digital
media demand a more reactive approach, while in Portugal, fact-checking faces resource constraints and the consolidation
of traditional media. ese dierences suggest that fact-checker strategies need to be adapted to each context, which also
explains variations in the use and eectiveness of technological tools.
In general, fact-checkers face specic challenges in the use of technological tools and in implementing ecient workows.
e survey revealed that the social networks most used by fact-checkers are X (formerly known as Twitter), WhatsApp and
Instagram. However, platforms such as WhatsApp and Telegram present signicant diculties due to the private nature of
communications, making it dicult to trace the source of disinformation. is problem highlights the need to develop more
advanced tools that allow fact-checkers to access and analyse content on these platforms more eectively.
e main tasks of fact-checkers include monitoring social media, searching for information and documentation, and writing
articles. It is notable that a considerable proportion of fact-checkers also participate in editorial meetings and use digital tools
to search for evidence. However, a small fraction of them engages in video editing and disinformation training, suggesting
potential areas of specialisation and professional development. e use of specialised tools such as Wayback Machine and
484 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Search by Image is common among fact-checkers, but there is a clear need for more training and resources in the use of
articial intelligence and other advanced technologies. Reliability, ease of use and time eciency are highly valued features
of these tools, underlining the importance of designing solutions that are not only eective, but also accessible and easy to
integrate into the daily workows of fact-checkers.
e results also indicate a signicant demand for additional training in AI techniques, OSINT (open-source intelligence), data
analytics and data visualisation. is need for training reects current gaps in the competencies of fact-checkers and highlights
the importance of ongoing educational programmes that address these critical areas. Integrating these competencies will
improve not only the accuracy and eciency of fact-checking, but also the ability of fact-checkers to adapt to the evolving
tactics of disinformation propagators.
While AI tools can facilitate the analysis and monitoring of large volumes of information, their implementation also presents
signicant challenges. In the US, PolitiFact has implemented natural language processing tools for automated detection of
claims to verify, with mixed results due to diculties in interpreting contextual nuances. In Latin America, Chequeado has
developed machine learning algorithms to analyse information on platforms such as WhatsApp, facing challenges related to
privacy and data access. ese examples show that while AI can improve verication eciency, its implementation requires
context-specic and context-tailored approaches.
AI’s ability to interpret cultural and linguistic nuances is limited (especially in languages other than English), and automated
tools can have insurmountable diculties in identifying disinformation in encrypted messages or on closed platforms. e
lack of transparency in the algorithms used to detect misinformation and the risk of bias in the results also raise ethical and
practical concerns. erefore, in its current state, AI should be seen as a complementary tool to human work, and not as a
substitute, in line with work that calls for human-in-the-loop approaches (Cantón-Correa, 2023; Das et al., 2023; Demartini
et al., 2020; La Barbera et al., 2022, 2022; Quelle & Bovet, 2024; Shabani et al., 2021; Yang et al., 2021). To improve their
eectiveness, it is crucial to develop technological solutions that can be more seamlessly integrated into the fact-checkers’
workows, while respecting privacy and legal regulations.
e data obtained show a complex and challenging picture for fact-checkers in the Iberian Peninsula. While technological tools
are available, their eectiveness depends mostly on the integration of human resources and adequate training. Addressing the
identied needs in terms of tools and training will be crucial to strengthen the capacities of information providers and improve
the ght against disinformation in the region.
5.1. Limitations of the study
e present study has several limitations that should be considered when interpreting the results obtained. Firstly, the low
response rate in the surveys, which was 10%, introduces a non-response bias that aects the external validity of the ndings.
is means that the results cannot be generalised to the entire population of fact-checkers in the Iberian Peninsula and Ibero-
America, as respondents may not adequately reect the diversity and needs of the entire population. Despite eorts to improve
the participation rate, such as sending out surveys at dierent times and contacting fact-checkers through dierent platforms,
workload and lack of time to participate in external research signicantly limited data collection.
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 485
Another signicant limitation is the exploratory nature of the participant observation conducted in the fact-checkers’
newsrooms. Although this technique allowed for a preliminary insight into the internal dynamics and methodologies
employed, time and access restrictions prevented a deeper and more systematic immersion. Furthermore, it was not possible
to access the fact-checkers’ oces in Portugal, which limited the ability to obtain a comprehensive perspective of the Iberian
verication ecosystem. ese circumstances reduced the ability to collect robust empirical data, and the ndings derived from
the observation should be interpreted as preliminary approximations and not as conclusive results.
Finally, limitations were identied in the implementation of articial intelligence (AI) technologies in fact-checking. Although
the available tools and their potential uses were explored, the eectiveness of these technologies is conditioned by factors
such as the ability of fact-checkers to use such tools eciently and the existence of a favourable regulatory and cultural
environment for their adoption. Dierences in the acceptance and use of AI in dierent contexts, as well as the ethical and
technical challenges associated with its implementation, limit the applicability of the ndings to other regions and suggest the
need to adapt international best practices to local specicities.
In future research, it is recommended to expand the sample of participants to obtain more representative results and to
deepen the use of participant observation with a more systematic and longitudinal approach. In addition, active collaboration
with fact-checkers to improve participation in surveys and the development of comparative studies across dierent contexts
can provide a more complete understanding of the challenges and opportunities in the use of advanced technologies in
fact-checking.
6. Conclusions
is research has provided a detailed and comprehensive analysis of the needs and challenges faced by information verication
professionals, also known as fact-checkers. rough a combination of quantitative and qualitative methodologies, the current
state of these professionals, their tools, workows and training needs have been explored. is in-depth analysis is essential
to understand and address the training, technological and practical gaps faced by fact-checkers, enabling the development of
more eective solutions tailored to their specic needs.
Survey responses reveal several key areas where fact-checkers need additional support or specialised tools. ese include
social media, image and video analysis and the use of articial intelligence (AI). In addition, signicant demand has been
identied for training in the use of AI, open-source intelligence (OSINT) techniques, data and information analysis, image
analysis and data visualisation. ese ndings underline the need to develop and provide access to specialised tools and
training that address these specic aspects, thus facilitating the work of fact-checking.
In addition, the report highlights the importance of trust, ease of use and time eciency as crucial features for the tools used
by fact-checkers. While other aspects such as privacy, user interface and price are also important, they are not considered
as crucial as the above. is nding suggests that, when using a tool, If users nd it easy and eective, they will be willing to
compromise on other aspects.
486 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
e nal considerations of this research highlight the need to address the gaps and challenges identied by developing tools
and providing training that eectively meet the needs of fact-checkers. As disinformation continues to evolve and spread
through various channels, it is imperative to equip fact-checkers with the necessary skills and tools to eectively combat
this phenomenon. With a proactive approach and tailored responses, the ght against disinformation can be signicantly
strengthened, thus supporting the integrity and veracity of information in society.
6.1. Constraints and needs of fact-checkers
e survey responses highlight several important shortcomings and limitations in the work of fact-checkers. One of the main
diculties identied is the lack of specialised and ecient tools for the analysis of social networks, images and videos, as well
as for the implementation of Articial Intelligence (AI) techniques in the verication process (Montoro-Montarroso et al.,
2023).
In addition, participants expressed the need for additional and specialised training in emerging and critical areas, such as the
use of AI, open-source intelligence techniques (OSINT) and data and image analysis. e demand for training in these areas
reects a skills gap that, if properly addressed, can empower fact-checkers to navigate and analyse the information landscape
more eectively and accurately. Training in these specialised skills will not only improve the quality of fact-checking work, but
also strengthen the resilience of practitioners in the face of changing and sophisticated tactics employed by the propagators
of disinformation.
Another identied constraint is the lack of access to reliable and rapid sources of information. Fact-checkers often have
diculty obtaining ocial data or responses from government entities and organisations, which delays and complicates the
verication process. In addition, the need for an accessible and reliable database of experts who can provide quick information
and clarication to facilitate the verication work was mentioned.
Finally, a lack in the number of editors specialising in specic topics, such as climate change, was noted. Subject-specic
expertise is vital for understanding and analysing information and disinformation in these elds, underlining the need to
invest in training and recruitment in specic subject areas. Overall, addressing these gaps and constraints will be critical to
improve the eectiveness and eciency of fact-checking work and ultimately to strengthen the ght against disinformation
in society.
e needs of fact-checkers are multi-faceted and include advanced technological tools, ongoing specialised training, access
to reliable sources of information and subject matter expertise. Addressing these needs is essential to empower fact-checkers
in their ght against disinformation, enabling them to work more eectively and accurately in an increasingly complex and
challenging digital environment.
6.2. Proposals and integrations
In view of the needs and gaps identied in the report, a few strategic initiatives are proposed to strengthen the professional
eld of verication. Firstly, it is imperative to develop advanced fact-checker tools and facilitate access to them. ese tools
must be able to eectively analyse social networks, images and videos, and integrate advanced Articial Intelligence (AI)
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 487
techniques. e creation of integrated platforms that consolidate several fact-checker tools into a single workspace can also
facilitate a smoother and more ecient verication process.
In addition, the implementation of continuous and specialised training programmes for fact-checkers is suggested. ese
programmes should cover key areas identied by practitioners, such as the use of AI, OSINT techniques, data and image
analysis and data visualisation. Training should be exible, accessible and in line with new trends and challenges in the eld
of information verication.
To cater for the need of a shared, fast and reliable access to information, we propose the creation of a centralised and accessible
repository that includes:
–
A database of fact-checkers already carried out by entities, including not only the reporting and publication of these, but also
the original complaint, audio-visual materials and everything that has served to contextualise the research.
–
A list of reliable sources of information, such as pages to ocial statistics and compilations of reports, classied by topic.
–
A directory of experts in various elds, institutions and countries needed for fact-checkers.
is resource could act as a unied and useful quick-reference directory for fact-checkers, facilitating access to available and
relevant knowledge in real time. Eorts that have been made in this direction, such as TrulyMedia, have not fully worked. In
addition, it is also recommended that clear and eective protocols for data analysis be established to guide practitioners in the
fact-checking process and help them navigate diverse and complex information sources.
Suggested proposals and integrations to improve the professional eld of verication include design thinking workshops that
directly involve fact-checkers. e aim of these workshops would be to co-create tools that t their specic needs, drawing
on their experience and direct knowledge of the eld. e active collaboration of fact-checkers in the design process not only
ensures that solutions are practical and relevant, but also encourages adoption and engagement with the tools developed. is
practice will identify emerging trends and changes in the needs and preferences of the fact-checkers, as well as evaluate the
eectiveness of the tools and training applied. is dynamic and adaptive approach ensures that fact-checker strategies keep
up with the fast pace of disinformation and technological innovations in the eld of fact-checking.
Finally, given the diversity of topics and the specialisation required in the fact-checking process, the incorporation and
training of editors specialising in key areas is recommended. e implementation of these proposals and integrations will not
only respond to the current needs of fact-checkers but will also strengthen the overall fact-checking infrastructure, making the
process more robust, reliable and better equipped to face future challenges in the ght against disinformation.
7. Acknowledgements
is article has been translated into English by María Delgado Mena to whom we are grateful for her work.
is work was supported by the IBERIFIER Plus project, co-funded by the European Commission under the Call DIGITAL-
2023-DEPLOY-04, European Digital Media observatory (EDMO) – National and multinational hubs, grant number: IBERIFIER
Plus – 101158511.
488 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
8. Specic contributions of each author
Name and Surname
Conception and design of the work
Javier Cantón-Correa, Andrés Montoro- Montarroso,
Juan Gómez-Romero and Miguel Molina-Solana
Methodology
Javier Cantón-Correa
Data collection and analysisJavier Cantón-Correa
Discussion and conclusionsJavier Cantón-Correa, Juan Gómez-Romero and Miguel Molina-Solana
Drafting, formatting, version review and approval
Javier Cantón, Andrés Montoro- Montarroso, Juan Gómez-Romero and
Miguel Molina-Solana
9. Conict of interest
e authors declare that there is no conict of interest contained in this article.
10. Bibliographical references
Agarwal, B., Agarwal, A., Harjule, P., & Rahman, A. (2023). Understanding the intent behind sharing misinformation on social media.
Journal of Experimental & eoretical Articial Intelligence, 35
(4), 573–587.
https://doi.org/10.1080/0952813X.2021.1960637
Alfonso, I. B., Domínguez, M. Á. C., & Caballero, R. R. (2021). El fact-checking como estrategia global para contener la
desinformación.
Estudios sobre el mensaje periodístico, 27
, 779–791.
https://doi.org/10.5209/esmp.76189
Angrosino, M. (2012).
Etnografía y observación participante en investigación cualitativa
(Vol. 3). Ediciones Morata.
Arcila, C. A., Alonso, M. O., & Herrero, D. B. (2020). Digitalización del periodismo iberoamericano y su efecto en la relación del
periodista con la audiencia. Estudio comparativo de Argentina, Brasil, Chile, Colombia, Ecuador, El Salvador, España, México y
Portugal.
Journal of Iberian and Latin American Research, 26
(3), 284–300.
https://doi.org/10.1080/13260219.2020.1909831
Berganza, R., Lavín, E., & Piñeiro-Naval, V. (2017). Spanish journalists’ perception about their professional roles.
Comunicar,
25
(51), 83–92.
https://doi.org/10.3916/C51-2017-08
Bericat, E. (1998).
La integración de los métodos cuantitativo y cualitativo en la investigación social: Signicado y medida
. Ariel.
Bontridder, N., & Poullet, Y. (2021). e role of articial intelligence in disinformation.
Data & Policy, 3
, e32.
https://doi.org/10.1017/dap.2021.20
Cantón-Correa, J. (2023).
SoundHive: Una propuesta de herramienta de monitorización para vericadores de información.
https://doi.org/10.5281/zenodo.10015755
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 489
Cantón-Correa, J., Sidorenko-Bautista, P., & Alba-Ruiz, R. (2024). Monitorización y detección automática de noticias mediante
inteligencia articial: Una propuesta de herramienta para redes sociales y plataformas.
Espejo de Monografías de Comunicación
Social, 25
, Artículo 25.
https://doi.org/10.52495/c1.emcs.25.p108
Cifuentes Arias, C., & Sixto-García, J. (2022). Competencias y perl profesional de los gestores de medios sociales en los medios
de comunicación: Detección de una fuerte apuesta por la automatización robótica.
Revista Latina de Comunicación Social, 80
,
1–23.
https://doi.org/10.4185/RLCS-2022-1544
D’Ancona, M. Á. C. (1996).
Metodología cuantitativa: Estrategias y técnicas de investigación social: 1
.
Das, A., Liu, H., Kovatchev, V., & Lease, M. (2023). e state of human-centered NLP technology for fact-checking.
Information
Processing & Management, 60
(2), 103219.
https://doi.org/10.1016/j.ipm.2022.103219
Demartini, G., Mizzaro, S., & Spina, D. (2020). Human-in-the-loop Articial Intelligence for Fighting Online Misinformation:
Challenges and Opportunities.
IEEE Data Eng. Bull., 43
(3), 65–74.
Echevarría, B. (2016). Más ‘fact-checking’ contra la posverdad.
Cuadernos de periodistas, 33
, 9–16.
Emerson, R. M., Fretz, R. I., & Shaw, L. L. (2011).
Writing ethnographic eldnotes
. University of Chicago Press.
https://doi.org/10.7208/chicago/9780226206868.001.0001
Esteban-Navarro, M.-Á., Nogales-Bocio, A.-I., García-Madurga, M.-Á., & Morte-Nadal, T. (2021). Spanish Fact-Checking Services:
An Approach to eir Business Models.
Publications, 9
(3), Artículo 3.
https://doi.org/10.3390/publications9030038
European Commission. (2018).
A multi-dimensional approach to disinformation: Report of the independent High level Group on
fake news and online disinformation
. Publications Oce.
https://data.europa.eu/doi/10.2759/739290
Fernández-Barrero, Á., Rivas-de-Roca, R., & Pérez-Curiel, C. (2024). Disinformation and Local Media in the Iberian Context: How
to Protect News Credibility.
Journalism and Media, 5
(1), Artículo 1.
https://doi.org/10.3390/journalmedia5010005
Flick, U. (2012).
Introducción a la investigación cualitativa
.
Flick, U. (2015).
El diseño de la investigación cualitativa
(Vol. 1). Ediciones Morata.
Groves, R. M. (2006). Nonresponse Rates and Nonresponse Bias in Household Surveys.
Public Opinion Quarterly, 70
(5), 646–675.
https://doi.org/10.1093/poq/n033
Hammersley, M., & Atkinson, P. (2019).
Ethnography: Principles in practice
. Routledge.
https://doi.org/10.4324/9781315146027
Herrero, E., & Herrera-Damas, S. (2021). El fact-checking hispanohablante: Competencias, dicultades y propuestas de mejora
desde la perspectiva de sus profesionales.
El Profesional de la Información, 30
(6).
https://doi.org/10.3145/epi.2021.nov.12
Kvale, S. (2012).
Las entrevistas en investigación cualitativa
(Vol. 2). Ediciones Morata.
La Barbera, D., Roitero, K., & Mizzaro, S. (2022). A hybrid human-in-the-loop framework for fact checking.
Proceedings of the
Sixth Workshop on Natural Language for Articial Intelligence (NL4AI 2022)
co-located with 21st International Conference of the
Italian Association for Articial Intelligence (AI*IA 2022).
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
490 |
nº 40, pp. 465-491
|
January-June of 2025
Proles and training needs of fact-checkers in the Iberian Peninsula: Impact of AI
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Lazo, C. M., Rodríguez Rodríguez, J. M., & Peñalva, S. (2020). Competencias digitales en periodismo. Revisión sistemática de
la literatura cientíca sobre nuevos perles profesionales del periodista.
Revista Latina, 75
, 53–68.
https://doi.org/10.4185/
RLCS-2020-1416
López-García, X., Rodríguez-Vázquez, A.-I., & Pereira-Fariña, X. (2017). Technological skills and new professional proles:
Present challenges for journalism.
Comunicar, 25
(53), 81–90.
https://doi.org/10.3916/C53-2017-08
Lotero-Echeverri, G., Romero-Rodríguez, L. M., & Pérez-Rodríguez, M. A. (2018). Fact-checking vs. Fake news: Periodismo de
conrmación como componente de la competencia mediática contra la desinformación.
Martín-Antoranz, P., Cabezuelo-Lorenzo, F., & Bartolomé-Muñoz-de-Luna, Á. (2019). Competencias y formación ética de los
profesionales de la comunicación corporativa en contextos interculturales.
Revista de Ciencias de la Comunicación e Información
,
59–72.
https://doi.org/10.35742/rcci.2019.24(1).59-72
Montoro-Montarroso, A., Cantón-Correa, J., Rosso, P., Chulvi, B., Panizo-Lledot, Á., Huertas-Tato, J., Calvo-Figueras, B.,
Rementeria, M. J., & Gómez-Romero, J. (2023). Fighting disinformation with articial intelligence: Fundamentals, advances and
challenges.
Profesional de La Información, 32
(3), Artículo 3.
https://doi.org/10.3145/epi.2023.may.22
Quelle, D., & Bovet, A. (2024). e perils and promises of fact-checking with large language models.
Frontiers in Articial
Intelligence, 7
, 1341697.
https://doi.org/10.3389/frai.2024.1341697
Quian, A. (2021). Observación participante en una organización de ltraciones periodísticas: El caso WikiLeaks.
Empiria. Revista
de metodología de ciencias sociales, 52
, 199–231.
https://doi.org/10.5944/empiria.52.2021.31370
Rodríguez Martínez, R., Mauri, M., Chaparro, M., Egaña, T., Fanals Gubau, L., Herrera, S., Morata, M., Moreno Gil, V., Luengo, M.,
& Palà Navarro, G. (2021). Desinformación y plataformas de fact-checking: Estado de la cuestión.
Rodríguez Pérez, C. (2020). Una reexión sobre la epistemología del fact-checking journalism: Retos y dilemas.
Revista de
comunicación, 19
(1), 243–258.
https://doi.org/10.26441/RC19.1-2020-A14
Sánchez-García, P., Merayo-Álvarez, N., Calvo-Barbero, C., & Diez-Gracia, A. (2023). Spanish technological development of
articial intelligence applied to journalism: Companies and tools for documentation, production and distribution of information.
Profesional de la información, 32
(2), Artículo 2.
https://doi.org/10.3145/epi.2023.mar.08
Santos, F. C. C. (2023). Articial Intelligence in Automated Detection of Disinformation: A ematic Analysis.
Journalism and
Media, 4
(2), Artículo 2.
https://doi.org/10.3390/journalmedia4020043
Shabani, S., Charlesworth, Z., Sokhn, M., & Schuldt, H. (2021). SAMS: Human-in-the-loop approach to combat the sharing of
digital misinformation.
CEUR Workshop Proceedings
.
Slapakova, L. (2021). Towards an AI-Based Counter-Disinformation Framework.
https://bit.ly/4fhbVmT
Spradley, J. (1980).
Participant Observation
. Holt, Rinehart and Winston, Inc.
Torcal, M. (2023).
De votantes a hooligans: La polarización política en España.
Catarata.
Ufarte-Ruiz, M.-J., Peralta García, L., & Murcia-Verdú, F.-J. (2018). Fact checking: Un nuevo desafío del periodismo.
El Profesional
de la Información
.
https://doi.org/10.3145/epi.2018.jul.02
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 40, pp. 465-491
January-June of 2025
Javier Cantón-Correa, Andrés Montoro-Montarroso, Juan Gómez-Romero and Miguel Molina-Solana
ISSN: 1696-019X / e-ISSN: 2386-3978
| 491
Vizoso, Á., & Vázquez-Herrero, J. (2019). Plataformas de fact-checking en español. Características, organización y método.
Communication & Society, 32
(1), 127–144.
Yang, J., Vega-Oliveros, D., Seibt, T., & Rocha, A. (2021). Scalable fact-checking with human-in-the-loop.
2021 IEEE International
Workshop on Information Forensics and Security (WIFS)
, 1–6.
https://doi.org/10.1109/WIFS53200.2021.9648388
Zhou, X., & Zafarani, R. (2021). A Survey of Fake News: Fundamental eories, Detection Methods, and Opportunities.
ACM
Computing Surveys, 53
(5), 1–40.
https://doi.org/10.1145/3395046
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>