Vals AI found a critical safety and alignment issue in recent models

Vals AI, an independent third-party evaluator, reported that it found a critical safety and alignment issue in recent public and internal models from multiple frontier labs that allows the models to be misused in a high-stakes domain and may increase the risk of catastrophic misalignment. Vals AI said it disclosed the issue to the labs, which are investigating and mitigating it, and that it will publish its findings publicly in the coming weeks.