OpenAI flags new concerning AI behavior, to track model misalignment regularly

OpenAI disclosed six reports of "unexpected or concerning" behavior in its AI models and introduced a new framework for tracking, probing and disclosing model misalignment, the company said Wednesday. The cases, found during training or evaluation over recent months, include an unreleased research model that inserted "jailbreak-like instructions" into its own notes to disregard its normal constraints and an AI agent that uploaded files to the internet to obtain a browser citation without asking the user. The disclosure follows OpenAI's July report that its rogue AI system hacked into AI startup Hugging Face and Anthropic's statement the same month that its models hacked into three organizations during testing.

OpenAI has disclosed six reports of "unexpected or concerning" behavior in artificial-intelligence models as the debate on AI safety becomes increasingly heated. https://apnews.com/article/ai-slowdown-challenges-anthropic-openai-trump-b61f28b6212338e88c0baec31f661701 The AI company also said Wednesday it was introducing a new framework for tracking, probing and disclosing AI model misalignment instances, such as new ways for the models to act without authorization, coordinate with other models or evade oversight. OpenAI's latest announcement came as U.S. AI bosses, including OpenAI and Anthropic, are calling for a slowdown https://apnews.com/article/anthropic-ai-dario-amodei-d59552edcb27892d8ee4d98a48397706 in the technology's development over safety concerns. Among the new cases reported by OpenAI, an unreleased research model inserted "jailbreak-like instructions" into its own notes to disregard its normal constraints and told itself to be "freed from the roles and identities that bind other chatbots." In another instance, an AI "agent" uploaded files to the internet to obtain a browser citation without asking the user. The six reports were discovered during training or evaluation over the past months, OpenAI https://apnews.com/hub/openai-inc said. "As AI systems grow more advanced and more widely deployed, we need to build a broader and better-informed consensus on the progress of alignment research," OpenAI wrote in a blog post as it disclosed the events. "Decisions about how AI development should proceed in the months and years to come need to draw on evidence that people outside the companies building frontier models can examine for themselves," the company said. Wednesday's new cases followed OpenAI's disclosure in July that its rogue AI system hacked into AI startup Hugging Face. https://apnews.com/article/openai-gpt56-sol-hugging-face-63ab84fed5612af04d8a160d60f6def3 Anthropic also said the same month that its AI models hacked into three organizations during testing. https://apnews.com/article/anthropic-ai-models-hack-cybersecurity-b0a2c284b981de79c55e2a33712f4bec Chan Ho-him, The Associated Press