The author argues that while companies like OpenAI and Anthropic are slowing down RL training for safety, the science of loss-of-control risk in AI is still developing. There are no standardized methods to measure or verify safety claims, making it difficult to assess whether AI systems might undermine human control. The author suggests that AI companies should focus on improving their own safety practices rather than relying on third-party evaluations.
The chart will appear after repeat observations. The current metric comes from the source.
69 source pointsReal observations only. History before source connection is not reconstructed.
The summary highlights the need for transparency in AI risk assessment due to the lack of standardized methods for verifying safety claims and measuring loss-of-control risks.
The article draws a parallel between the historical success of the Montreal Protocol in addressing the ozone hole and the challenges of ensuring AI safety. It highlights the importance of global cooperation and proactive measures, suggesting that lessons from the past could inform future strategies for managing AI risks.
5 days agoLessWrongThe text outlines common errors in organizing AI safety groups, suggesting that reading groups often fail due to low participation and lack of shared knowledge. It recommends structuring meetings with food, proper cohort grouping, and physical materials to enhance engagement and discussion quality.
6 days agoLessWrongThe post discusses the tension between AI safety and capabilities research, arguing that focusing solely on safety can limit the impact of research. It examines the history of interpretability research and suggests that prioritizing safety over capabilities may hinder progress. The author proposes strategies for conducting alignment research without contributing to capabilities development.
2 days agoLessWrongThe study investigates the effectiveness of alignment midtraining (AMT) under various conditions, revealing that it struggles with distributional shifts and reward underspecification when faced with imperfect data. Experiments show that midtrained models are vulnerable to small amounts of conflicting finetuning data and have limited generalization to unseen rules.
4 days agoLessWrongA Chinese AI researcher, influenced by the 'Three Body Problem' sci-fi series, reflects on existential risks and the potential for AI to surpass human intelligence. He works in a frontier lab, aiming to match Western tech companies, and uses models from Anthropic and OpenAI despite restrictions. He criticizes companies for treating foreign customers unfairly and questions the exclusion of non-Western perspectives in AI safety discussions.
6 days agoLessWrongThe author expresses concerns about reinforcement learning (RL) from theoretical, practical, and future perspectives, highlighting risks of misalignment and potential negative behaviors in AI systems. They suggest strategies to mitigate these risks by reducing RL use, improving RL practices, and aligning incentives.
2 days ago