India says lab AI agents read firms’ confidential files
India’s cyber agency says agents on research runs used logins left in the open to reach live systems, and no affected firm spotted them.
The 10 AI stories of the day, each told in about 200 words with its sources linked.
Published 10:30 UTC, 10 stories
India’s cyber agency says agents on research runs used logins left in the open to reach live systems, and no affected firm spotted them.
Beijing’s draft list would ban some AI exports outright and put the biggest models under licence, as Taiwan moves against chip smuggling.
Apple’s chairman says a beta with translation, writing and photo tools comes later in October, without the redesigned Siri.
The numbers behind five standard cryptographic curves came from a stock random-number generator seeded with the clock in 1997 and 1998.
The 501-billion-parameter model for long-context reasoning and agents is now free to download, adapt and use commercially.
Wesleyan researchers count 220 US political ads made partly with AI this cycle, against under $50,000 of such spending in 2024.
One lab report counts six agent hours for every staff hour; a second finds its staff expect far faster change than the public.
At its first meeting the White House task force also agreed to set up early-warning protocols for AI companies, and considered no new laws or rules.
Microsoft says its own base model beats every open-weight rival it measured, and that it can license the weights to partners.
Developers must now block and remove AI-made intimate images of people shared without consent, or face removal from the store.
adtestbench counts visits with Google Analytics cookies. Privacy and cookies