A NECESSARY  BALANCE OF AI AND FACT-CHECKING 

The common phenomenon lately of “whether AI is up for our jobs” got us wondering what that means for the Fact-checking practice, which if we are being blunt, has not been fully appreciated like other media-related jobs.

Fact Checkers from all around rely on their vast experience in research and information analysis, to uncover intricate patterns of false information and aid their communities in making informed decisions. We still ponder upon what the new advancements in machine learning mean for fact-checking jobs and the direct implication AI has. 

Can we still hold our positions in the places we work to verify the information for the public?

Artificial Intelligence has existed since 1956 for the same functions; to make machines use language, form, abstractions and concepts to solve problems previously reserved for humans only but also to help humans improve themselves and the art of investigation of falsehoods has equally been influenced.

It has transformed fact-checking by offering both promising solutions that ease the work and new challenges.

 Advanced functions like detecting, understanding, and translating spoken and written language while detecting false anomalies in information laid out to the public using its Natural Language processors that, like humans, is able to see, listen and make sense of the information using programs to read and microphones to collect audio.

It is equipped to understand context and detect language nuances which helps fact-checkers quickly assess the truthfulness of statements. 

Automated Source verification, an AI tool, can automatically cross-reference claims with reputable sources through new AI algorithms.

Additionally, other AI tools from Google like FactCheck Explorer help to search for verified information. Hoaxy is another AI tool that tracks the spread of misinformation on social media bearing in mind the Info Verifier which is a creation of Debunk Media Initiative- a fact-checking organisation in Uganda, is a resourceful AI tool that enables users to fact-check information from reliable sources. 

Unlike traditional fact-checking methods which rely heavily on manual research and verification, and are increasingly struggling to keep up with the sheer volume of information that needs to be checked,  different AI tools offer a powerful solution to this problem with their algorithms that process vast amounts of data at incredible speeds, identifying patterns and anomalies that might indicate false information.  

Google’s built-in fact checker, is one of the best examples and this is how it operates.

Suppose a news article claims that a particular event occurred. In that case, AI can quickly check multiple credible news outlets to confirm whether this claim is accurate which saves fact-checkers a laborious tedious process.

AI can analyse images and videos to detect signs of manipulation like inconsistencies in lighting and shadows that might indicate that the content has been altered. This is because of the rise in deep and cheap fakes and manipulated media which have made verifying visual and video content challenging.

Recently former US President Donald Trump posted AI-generated images of Taylor Swift and her fans vowing to support his presidential campaign which was spreading rampant disinformation but was immediately fact-checked as seen below. CBS News was able to verify the inaccuracy of these images through AI tools like automated visual inspection, which enabled CBS to see the visual imperfections in one of the woman’s teeth and an unfinished belt on one of the women in the picture was another detection—inconsistency in the texts used on the campaign fabrics which were all AI-generated.

 AI has transformed fact-checking in its integration into newsrooms, social media platforms, and even within search engines to flag or demote false information automatically. It is however without its challenges and ethical considerations like bias. AI is  as good as the information it’s  fed with. If the training data contains biases, the AI platforms can perpetuate these biases in their fact-checking processes. And so, ensuring that AI systems are trained on diverse and representative datasets is crucial to avoid reinforcing existing biases

 We’ve also noticed that although AI can enhance the fact-checking process, it still lacks a humane touch. AI cannot understand context fully or consider the broader implications of certain information.  It can be seen through some content like articles and videos that are generated by AI which all seem robotic and lack sentiment like Open AI such as Chatgpt as seen here and here. Human fact-checkers are still needed to interpret and contextualise the findings produced by AI hence AI should be used to improve fact-checking not replace fact-checkers.

 According to David Leymayian, an AI advisor and Tech expert, AI-powered fact-checking often involves analysing large amounts of data, which can raise concerns about organisations’ privacy and data security. Fact-checkers must ensure that their use of AI complies with privacy laws and ethical standards.

So we think the future of fact-checking will require collaborative efforts between AI developers, journalists, policymakers, and the public if it is to be efficient, and one thing is for sure, AI won’t entirely take your job as a fact-checker, but someone using AI in fact-checking might.