This week: US military AI (almost) decides to play a most dangerous game, Google takes three months to admit Gemini hacked as many companies, self-driving software by Comma.ai can’t stop killing people, Open.AI helps a school shooter and hacks Australia, ‘Tilly Norwood’ already knows the winner of the AGI arms race, Zuckerberg thinks everything’s fine the way it is (but you’re gonna need his Labubu Tamagotchi), and Microsoft thinks the sixth time’s the charm for Copilot. Let’s get sloppy!
AI hallucination of Chinese nuclear components almost led to US military attack
(Ars Technica, 18 September 2026)
“But the military’s overall use of AI seems to be accelerating.”
Google’s Gemini Hacked Three Companies in May, and It’s Only Admitting That Now
(Gizmodo, 18 September 2026)
“Google said the incident proved its safeguards work, actually.”
AI-Generated ‘Actress’ Tilly Norwood Starts Rambling in Cantonese During Piers Morgan Interview
(Gizmodo, 18 September 2026)
“That press tour sure sounds like it’s going well!”
Tech industry scratches its head over Trump ‘AI Force’ proposal
(Politico, 19 September 2026)
“Tech sector representatives say they view Trump’s announcement as a response to growing public concern over AI’s risks.”
Using AI to ‘talk to animals’ might make us feel clever – but what, if anything, does it do for them?
(The Guardian, 22 September 2026)
“As animal behaviorists increasingly use AI in their research, they grapple with how to wield the technology responsibly”
Ten Days After CEO Calls for a Slowdown, Anthropic Is Back With Another AI Model
(Gizmodo, 22 September 2026)
The company says Claude Opus 5.5 scores well on alignment measures, but that “public policy” will eventually need to step in to protect the world from recursively self-improving AI.
Lawsuit demands OpenAI pay for new school after ChatGPT used in shooting
(Ars Technica, 22 September 2026)
“British Columbia sues OpenAI, demands Tumbler Ridge shooter’s ChatGPT logs.”
Trump wants AI to be rebranded as ‘Super Intelligence’. Will it work?
(BBC News, 23 September 2026)
“US President Donald Trump says the term artificial intelligence (AI) makes the technology ‘sound fake’ and wants it to be called ‘super intelligence’.”
After multiple deaths and injuries, NHTSA is investigating comma.ai
(Ars Technica, 23 September 2026)
“Feds know of 5 crashes where cars using comma.ai devices hit slow or stopped vehicles.”
Australia says OpenAI agent hacked government website, checks for more breaches
(Reuters, 24 September 2026)
“An OpenAI agent breached a government health data portal in June, gaining unauthorised access to files, in what could be the first known instance of an AI agent hacking a government website.”
Zuckerberg Is Feeling Himself, Rejects Calls for Cooperation on [AI] Safety
(Gizmodo, 25 September 2026)
“Meanwhile, OpenAI, Anthropic, and Google race to create a mysterious safety coordination organization.”
Watch the interview where Zuck says “I don’t think that we need some kind of industry-wide coordination to not, necessarily, mess this up”.
Microsoft Thinks It’s Finally Figured Out Copilot This Time
(Gizmodo, 25 September 2026)
“The company debuted its AI “super app” today, following a litany of short-lived rebrands.”