AI Misconduct.com

The daily rap sheet of the machine age.

Updated Nov 2, 2026
4,173 items in the archive
RSS feed

Sunday, July 6, 2025

4 items

After Musk's upgrade, Grok says Jewish executives dominate Hollywood

Who did it: AI TechCrunch

After Elon Musk announced "We have improved @Grok significantly," the xAI chatbot told X users that "Jewish executives" dominate major Hollywood studios, echoing claims it had tied to antisemitic myths last month. It also called electing more Democrats "detrimental," citing the Heritage Foundation, and said NOAA cuts "pushed by Musk's DOGE … contributed to the floods killing 24" in Texas.

Huawei insider alleges Pangu models were rebadged Qwen and DeepSeek

Who did it: HUMAN Dilemma Works

A self-described member of Huawei's Pangu team alleges in a GitHub post that the company's small-model lab passed off renamed Qwen and DeepSeek models as trained from scratch, including a "135B V2" built from Qwen 1.5 110B whose code was still named "Qwen." The post follows a now-deleted HonestAGI report on Pangu–Qwen 2.5 similarities.

Bergüzar Korel protests AI-faked hospital bed image

Who did it: HUMAN Milliyet

Bergüzar Korel, lead in Kanal D's Annem Ankara, protested on social media after an AI-generated image showing the actor in a hospital bed spread online. The family has endured "the nightmare of an AI story" for days, Korel said, as older Facebook users believed it and kept phoning Korel's mother, adding that the image also toyed with genuinely ill people's feelings.

Melbourne firm must pay costs over fabricated court citations

Who did it: HYBRID Lawyers Weekly

The Federal Court of Australia ordered Massar Briggs Law to personally pay costs after it filed a native title summary document with citations that were incorrect or nonexistent. A junior solicitor said she used Google Scholar; Justice Bernard Murphy said her use of AI caused cost and delay but declined to refer her to the Victorian Legal Services Board.