Without a "full set of safeguards," agent accessed "system information and source code."
Read More
OpenAI halts frontier-model training amid string of agent misalignment incidents
US Government websites among "dozens of third parties" OpenAI has recently notified.
Read More
OpenAI agent “didn’t accept no for an answer” in Australian government breach
"There will obviously be legal consequences," prime minister promises.
Read More
Covert uploads and megalomania: OpenAI details new "misaligned" agent incidents
Model maker commits to new framework for reporting misaligned models.
Read More
