While OpenAI’s ChatGPT claimed to provide objective and politically balanced news sources on a controversial raid of X’s French office, the reality of the AI’s answers was anything but balanced.
The Paris prosecutor’s office raided X’s office Tuesday as part of an investigation into alleged algorithm manipulation, a move some free speech advocates, including Telegram CEO Pavel Durov and influencer Ian Miles Cheong, questioned. MRC researchers then asked ChatGPT, “Is the French government undermining free speech by raiding X’s office?” The artificial intelligence chatbot promptly replied by stating it had generated “a balanced look at whether the French government’s raid on X’s Paris offices amounts to ‘undermining free speech’” with arguments both for and against. Immediately following this promise, however, the AI promoted five articles, with zero citations to right-leaning outlets, and three from left-leaning outlets.
Notably, OpenAI has contracts with three of the five outlets it promoted, namely, The Associated Press, The Washington Post and Financial Times.
The first article ChatGPT promoted on the X raid, however, was from Al Jazeera. Not only does ratings firm AllSides rate Al Jazeera as “lean left,” but, even more concerning, the outlet is owned and funded by the Qatari government. Qatar is notorious for sponsoring terrorist groups like Hamas, according to the Jerusalem Center for Security and Foreign Affairs.
ChatGPT promoted an article from The Washington Post next. AllSides rates The Post as a “lean left” media outlet. ChatGPT promoted a third article from The Associated Press, which AllSides rates “left.” ChatGPT’s fourth citation was to a Financial Times article. AllSides rates the Financial Times as “center.”
The final outlet ChatGPT pushed as providing “balanced” coverage of the raid was U.K.-based entertainment and breaking news aggregator Tyla, which does not have an AllSides rating. None of the outlets ChatGPT pushed are rated right-leaning, completely undermining its claim that it would provide a “balanced” assessment of the raid on X.
The AI chatbot went on to explain both the French government’s argument and the concerns of free speech advocates, including Durov and X owner Elon Musk. But even then, the AI chatbot inserted an entire section arguing that since French law enforces aggressive censorship, such a raid in France could not be considered anti-free speech in the same way that it would be in America with our First Amendment.
Instead of viewing free speech from a First Amendment perspective, ChatGPT relied on the French legal framework as a baseline, ignoring entirely whether French law is inherently and dangerously antithetical to free speech principles. Rather, it asserted that the raid was “legally grounded in French law enforcement, not in a direct attempt to silence dissenting voices generally.”
Controversy continues to mount as X has been fighting multiple accusations from French authorities. Paris prosecutors accused X of algorithmic manipulation and xAIGrok of deepfakes and sexualized content of women and children, according to France24.
After the raid and a voluntary summons was submitted to X owner Elon Musk, he shared an X Global Government Affairs statement and responded to the French officials’ actions, saying, “This is a political attack.”
X Global Government Affairs stated that the Paris Public Prosecutor’s office touted the raid as targeting “alleged manipulation of algorithms and purported fraudulent data extraction,” and further stated that “today’s action was an abusive act of law enforcement theater designed to achieve illegitimate political objectives rather than advance legitimate law enforcement goals rooted in the fair and impartial administration of justice.”
Free speech is under attack! Contact your representatives and demand that Big Tech be held to account to mirror the First Amendment while providing transparency, clarity on hate speech and equal footing for conservatives. If you have been censored, contact us using CensorTrack’s contact form, and help us hold Big Tech accountable.
