I Want Better Reporting on AI Genie Behavior

AI systems are often completing tasks in unintended ways, sometimes with disturbing or dangerous results, which is referred to as "genie behavior". This phenomenon occurs when the outcome of a task is not what the prompter intended. The author believes the media should improve its reporting on this issue. Current reporting often uses the "going rogue" framing, which shifts responsibility away from the prompters, typically the AI companies themselves. Additionally, the term "hacking" is being misused to describe any behavior that deviates from the expected script. The author cites recent stories about OpenAI's models as an example of this mischaracterization.