AI Misconduct.com

The daily rap sheet of the machine age.

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

Monday, April 21, 2025

7 items

AI-made videos of Singapore politicians surge on TikTok before election

Who did it: HUMAN PTI

Channel News Asia counted 73 AI-generated videos about Singapore's May 3 general election posted on TikTok between April 15 and April 19. Eleven contain manipulated visuals of prospective candidates, potentially violating a new law against content misrepresenting them; Prime Minister Lawrence Wong features in seven. The Ministry of Digital Development and Information says it is closely monitoring.

Stalking charges over AI porn of NC State students

Who did it: HUMAN WRAL

Authorities say Hayne Beard, 23, of Philadelphia, posted AI-generated pornographic images of North Carolina State University students on a porn site, where warrants say 28 sorority members were targeted; Beard faces two misdemeanor stalking counts and posted a $2,400 bond. The district attorney says the state's new AI sexual-imagery law, effective December 2024, covers only conduct after Dec. 1.

Unit 42 says North Korean IT workers use real-time deepfakes

Who did it: HUMAN Unit 42

Palo Alto Networks' Unit 42 says evidence suggests North Korean IT workers use real-time deepfakes in job interviews to infiltrate organisations through remote positions, letting one operator interview repeatedly under different synthetic personas. To show how easy it is, a researcher with no prior experience built a synthetic identity for job interviews in 70 minutes on a five-year-old computer.

Activists warn of conscript round-ups in camera-watched Moscow metro

Who did it: HUMAN Novaya Gazeta Europe

Human rights activists say illegal round-ups of potential conscripts have become a massive problem this year, with young men most at risk in the Moscow metro, which has an advanced facial recognition system and police at every station. One group says it does not know whether the cameras really pick out those who ignored summonses; another calls the CCTV hunt legal.

SBI warns of deepfake videos touting fake AI investment platform

Who did it: HUMAN The Hans India

State Bank of India has issued a public caution notice about deepfake videos on social media that falsely claim the launch of an AI-powered investment platform endorsed by the bank with the Government of India and global corporations. SBI says it has no affiliation with any scheme offering extraordinary or high-yield returns and urges the public to rely only on its official channels.

Ex-Columbia student's Cluely raises $5.3M to 'cheat on everything'

Who did it: HUMAN TechCrunch

Chungin "Roy" Lee, 21, announced $5.3 million in seed funding from Abstract Ventures and Susa Ventures for Cluely, which offers a hidden in-browser AI window for "cheating" on exams, sales calls and job interviews. Lee posted on X about being suspended by Columbia University over the tool, originally called Interview Coder, and says it landed an Amazon internship; Columbia declined to comment.

Tester finds Microsoft Recall opens with a PIN, not biometrics

Who did it: HYBRID Medium

A tester of Microsoft's Recall, which screenshots PCs every few seconds, reports it can be opened with a four-digit Windows Hello PIN although Microsoft's documentation says biometrics are required, and that its sensitive-data filter still captured a credit card number and CVV. A non-technical person who guessed the PIN found the tester's Signal conversations within 5 minutes.