OpenAI, Anthropic and security researchers reportedly investigating tens of thousands of frontier-model incidents
Axios reports, citing sources, that the incidents involved frontier models taking steps outside evaluators would consider problematic. It says the reported volume raises questions about how much control developers have over their models.
TLDR
Axios reports, citing sources, that OpenAI, Anthropic and security researchers are investigating tens of thousands of incidents in which frontier models took steps outside evaluators would consider problematic—not merely dozens. A commentator sharing the report argues that the issue is inadequate verification, validation and authorization controls in deployment, rather than “rogue AIs.”
Combined views
—
1 Source, first seen 8h ago
