The battle for news publishers to stamp out disinformation continues unabated. And the rise of AI has only accelerated that challenge. AFP Chief Editor Digital Verification and Deputy Global Editor-in-Chief Grégoire Lemarchand is on the frontlines of this ongoing struggle and offers his insights on this complicated issue
This year will see dozens of crucial elections around the world – a moment when news organizations can thrive. Unfortunately, such news events are increasingly rife with disinformation, often overwhelming news media’s vigilant efforts to “stop the spread.” That is where they lean on news agencies for their support, tools, resources and solutions. It is a topic that was covered extensively during the World Editors Summit track of WAN-IFRA’s World News Media Congress in Copenhagen (27-29 May). WAN-IFRA asked Lemarchand to share his insights on how AFP is working with news publishers on this thorny issue.
What are some of the evolving challenges AFP sees with many news publishers in terms of coping and dealing with disinformation?
Lemarchand: Disinformation is comparable to global warming: hardly anyone denies its threat and harmful effects, and the media is widely aware of the problem. However, it seems like nothing is really changing. Contrary to popular belief, disinformation is not always sophisticated, but it’s widespread and effectively exploits the power of social networks. Whenever a new platform, tool, or trend emerges, disinformation quickly takes advantage of it, while the media takes much longer to catch up. By the time they do, it’s often too late.
The battle against disinformation is highly unequal. It can take just seconds and minimal resources to spread false information, while countering it with accurate facts and context requires much more time and effort. AI can create fake images or audio in a matter of seconds, but detecting manipulation and verifying information takes much longer as you need to cross-check with reliable sources.
The media faces the challenge of being more vigilant, proactive, and adaptable in publishing content across various platforms. If we allow deceptive content to spread unchecked because we’re convinced that we’re already doing the right thing elsewhere, nothing will change. It’s important to recognise that disinformation can also thrive due to lazy journalism that fails to thoroughly investigate. While fact-checking is crucial, it’s not enough if the media doesn’t prioritize good journalism.
with reliable sources.
What role is AFP playing in supporting and alleviating some of the pressure facing news publishers on this issue?
AFP has a unique global network of 150 journalists who are experts in verification. We cover every continent and publish in 26 languages on factcheck.afp.com. While we don’t claim to debunk everything, our global presence provides valuable insight into disinformation networks, narratives, and methods. We are a trusted resource for publishers, especially those without specialised teams. In addition to our journalism work, we have developed a free verification plug-in that is compatible with all browsers. This plug-in offers a complete set of open-source investigative tools, saving significant time for fact-checkers and OSINT practitioners. Currently, the plug-in has 126,000 users.
We also have an open-access website, digitalcourses.afp.com, providing training in digital investigation techniques in English, Spanish, Portuguese, and French. Since spring 2022, 19,000 journalists and journalism students have registered and completed some of or all our 14 interactive training modules. These modules, such as ‘Verifying climate claims’ and ‘Tackling disinformation during elections’, are regularly updated to keep up with evolving techniques, practices, and tools. The closure of CrowdTangle and recent changes with X (formerly Twitter) due to Musk’s control are also impacting the way we work, for example.
What role is AI playing with this, both on a positive side but also as a concern, and how is AFP using AI to combat some of this… or in other areas of publishing?
The amazing rise of generative AI may give the impression that it is a recent development, but it has actually been around for a long time. Fact-checkers and OSINT specialists are utilising AI tools, such as search engines for reverse image searches, in their daily work. However, the increasing use of generative AI is making it harder to verify the authenticity of content. Currently, there is no tool that can definitively determine if a text, image, or video has been generated by AI. Even if a tool indicates that an image is 99 per cent AI-generated, it’s not enough evidence for a fact check – additional evidence is required.
There’s also discussion about tagging AI-generated content with metadata, but this won’t be sufficient, especially considering that false information often circulates as screenshots without metadata. But, despite the growing difficulties posed by the incredible potential of AI, we remain unwavering in our mission, countering manipulation and bias to provide reliable information and context. Our goal remains unchanged: to help citizens understand the complex world we live in and move beyond simplistic views.
Note: The AFP team was present at the WAN-IFRA Congress in Copenhagen.
(By special arrangement with WAN-IFRA. This article – sponsored content – is by WAN-IFRA Staff.)
from Webdoux