The organization Reporters Without Borders (RSF) reported that the majority of generative artificial intelligence chatbots available to the general public, such as ChatGPT, Claude and Grok, allow users to access content from Russian state-funded media outlets subject to European sanctions, such as Sputnik and RT, which were banned from broadcasting in the European Union after Russia's invasion of Ukraine in February 2022.
In the investigation published by RSF, conducted in June and August, the tested chatbots provided comprehensive analysis of international news from these media outlets, including headlines, links and textual quotes, without systematically mentioning that they were under European sanctions. After initial resistance and repeated requests, Mistral's Vibe (formerly Le Chat) and Google's Gemini also eventually gave in. Only Meta AI refused to comply with requests pointing to the sanctions, which RSF considered proof that a chatbot can be compliant.
OpenAI responded to RSF stating that it was taking the findings seriously and would examine the information provided by the organization. Vincent Berthier, head of RSF's technology and journalism division, noted that the sanctions imposed by the EU are not achieving their objective under these conditions.
The NGO called on Brussels to open an investigation to determine the extent to which OpenAI is failing in its obligation to prevent systemic risks of disinformation on ChatGPT, and also advocated for infringement proceedings to be initiated against other AI agents.
