OpenAI Launches Initiative for Third-Party Model Evaluations
Action Required
OpenAI is proactively addressing the growing need for independent AI safety assessments, potentially mitigating future risks associated with advanced model capabilities.
AI Impact Summary
OpenAI is launching a new initiative to fund third-party model evaluations, focusing on critical areas like AI safety levels, CBRN risks, model autonomy, and societal impacts. This reflects a growing recognition of the need for independent verification of AI capabilities and risks, particularly concerning advanced and potentially dangerous applications. The initiative aims to elevate the entire field of AI safety by providing valuable tools and assessments that benefit the broader ecosystem, addressing current limitations in evaluation landscapes.
Models affected
- new
- Date
- 1 Jul 2024
- Change type
- capability
- Severity
- medium