A driverless Waymo with no safety operator struck a child within two blocks of a Santa Monica elementary school during drop-off hours on January 23, 2026, after the child ran out from behind a double-parked SUV; Waymo reported minor injuries. NHTSA opened a preliminary evaluation on January 28 into whether the vehicle exercised appropriate caution near the school.
Amazon reported hundreds of thousands of pieces of suspected child sexual abuse material found in data gathered to train its AI models last year, most of the more than 1 million AI-related reports NCMEC received in 2025. Amazon says the data came from external sources and it lacks origin details; NCMEC says the reports are 'inactionable.'
Two security researchers found that AI dinosaur toy maker Bondu's web portal let anyone logging in with a Google account read more than 50,000 children's chat transcripts, plus names, birth dates and family member names. Bondu says fixes were completed within hours and it found no other access; Senator Maggie Hassan has demanded answers to 10 questions.
Defence Minister Shinjiro Koizumi warned on X on January 29 that several people had received suspicious calls from someone posing as the minister's secretary, some put through on WeChat video to a fake Koizumi whose image and voice were probably AI-generated and whose answers did not add up. Koizumi vows a strict response, including consulting police.
Ukraine's Centre for Countering Disinformation warned on Facebook of a new wave of AI-generated videos depicting supposed MPs, police officers, recruitment centre staff and Ukrainian soldiers as deliberately fat caricatures. The centre says the enemy mass-produces such content disguised as Ukrainian satire, aiming to cultivate hatred of the military, police and officials and destabilise Ukraine.
Who did it: HYBRID AI Hallucination Cases database
A Toronto Small Claims Court deputy judge ordered paralegal Derya Vural to personally pay a $2,000 penalty after her submissions cited a non-existent section 13 of the Statute of Frauds and cases such as 'Sandler v. Goldstein, 2007 ONCA 153' that could not be found. Vural denied using AI; the court said the citations had 'all of the hallmarks' of hallucinations.
Egyptian artist Yasmin Abdelaziz announced on Wednesday that she has filed lawsuits and official complaints against pages and accounts that published fabricated or offensive images of her, which the report says were faked using AI. What happened, she said, had "gone beyond the limits of silence", and she would no longer simply ignore such abuses.
