The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Researchers from Oxford and Potsdam found that AI language models used to generate and edit social media posts systematically introduce subtle biases, especially on controversial topics. These biases accumulate through millions of interactions, shaping public opinion and raising concerns about transparency and digital regulation.[AI generated]
Why's our monitor labelling this an incident or hazard?
The article explicitly involves AI systems (large language models) used to generate and edit social media content. The AI's use has directly led to subtle but systematic bias in content that influences public opinion, which constitutes harm to communities by manipulating discourse and potentially undermining democratic processes. The harm is realized and demonstrated through research and simulations, not merely potential. Therefore, this event meets the criteria for an AI Incident rather than a hazard or complementary information. It is not unrelated or beneficial use, as the AI is the source of the harm, not a countermeasure.[AI generated]