Post #2128165

@mimarek There's a ~15 minute interview with the paper's authors here: https://www.youtube.com/watch?v=zdIqsbtqhcA An interesting detail from their experiment is that they manipulated ChatGPT's responses to be less accurate in some cases. That might be a useful hiring tool in the age of #AI, where employees are expected to use AI, but you want to filter for the ones that understand or verify the output vs blindly trust it and adopt it as their own work.