The Core Contradiction
The fundamental paradox lies in AI's dual nature: it is both a powerful tool for creating sophisticated falsehoods and a promising ally in debunking them. Fact-checkers are swamped. The sheer volume and speed of online misinformation have made manual
verification methods increasingly challenging. AI offers the potential for speed and scale, helping to process vast amounts of content at a pace no human could match. Yet, this is the same technology that can 'hallucinate' facts, create fake images of world events, and generate persuasive but entirely baseless text, sometimes with little human supervision. This puts fact-checkers in the tricky position of using a tool they fundamentally cannot trust to make final judgements.
Not for Judgment, but for Legwork
So, how are professionals navigating this? The key is in carefully defining AI's role. Rather than asking AI to be the arbiter of truth, newsrooms and fact-checking organisations are using it as an assistant for laborious tasks. This includes transcribing interviews, translating content, and summarising long reports to quickly identify key claims that need verification. Some organisations are using AI tools to perform initial scans of social media to spot trending narratives or identify multiple instances of the same false claim spreading online. The Norwegian fact-checking organisation Faktisk, for instance, has used AI to create informative maps to support its investigations. The rule is clear: AI can do the legwork, but the final call on what is true or false remains a deeply human process, requiring context, nuance, and critical judgment.
Building the Guardrails
Recognising the risks, the journalism community is proactively building ethical frameworks. Organisations like the Poynter Institute and the International Fact-Checking Network (IFCN) have developed guidelines for the responsible use of AI. These principles emphasise that any output from an AI tool should be treated as unvetted source material that requires independent verification. Core tenets include transparency with the audience about when and how AI is used, and a non-negotiable requirement for human oversight on all published content. This isn't about blind adoption; it's about creating a system of checks and balances where technology serves, rather than subverts, journalistic integrity.
An Uneasy Alliance
The relationship between fact-checkers and AI is best described as an uneasy, evolving alliance. On one hand, AI can help level the playing field in an information war where bad actors are already using AI to generate propaganda and deepfakes at an alarming rate. On the other hand, over-reliance on these tools could dull critical thinking skills or introduce subtle biases embedded in the AI models themselves. A recent Penn State study found that people trust AI and human fact-checkers almost equally, but for different reasons; they trust AI for scale and pattern recognition, but humans for nuanced interpretation. This highlights the need for a hybrid approach, where the machine's ability to scan and sort is combined with a human's ability to understand intent and context.
