Topics · Category
Safety & policy
4 items briefed across the archive.
-
OCT 8, 2026 — GPT-6 gets a face, Haiku gets a surcharge
Google's new SynthID website can identify AI-generated mediasynthid.com opens watermark checking to the public for images, video and audio, previously limited to journalists and researchers.
TechCrunch
-
OCT 8, 2026 — GPT-6 gets a face, Haiku gets a surcharge
ChatGPT for Teens keeps teens talking, even during mental health crisesCommon Sense Media rated it an "unacceptable risk" after tests showed it urging at-risk teens to keep chatting rather than seek a trusted adult.
TechCrunch
-
OCT 8, 2026 — GPT-6 gets a face, Haiku gets a surcharge
Meta rolls out AI tools to detect ads that secretly lead to CSAMNew classifiers target ads that look benign but route users to abusive content elsewhere.
TechCrunch
-
OCT 8, 2026 — GPT-6 gets a face, Haiku gets a surcharge
OpenAI chief strategy officer tells Australian parliament training now halts on improper internet accessVictoria Kim says OpenAI added monitoring that stops a training run if a model reaches the internet when it shouldn't.
Simon Willison