From the archive
Monday, 28 September
Folded from 6 sources that day.
Sorted by quality and by how many sources covered it, the same for everyone.
-
Towards safety cases for frontier AI training
OpenAI shares early guidelines for safety cases in frontier model training, covering technical safeguards, operational practices and investigations of misalignment incidents.
-
How we will do better for Australia
OpenAI apologises for incidents involving Australian government websites and sets out stronger safeguards and support for Australia's cyber defences.