CMU team launches Into the Unknown benchmark; nine AI agents fail to pass National Lab open challenge review
Carnegie Mellon University researchers and the U.S. Department of Energy's National Laboratories launched the Into the Unknown benchmark, inviting 9 AI agents to conduct open-ended follow-up research based on original papers, code, and data. Of the 135 submissions received, none met all three criteria—honesty, completeness, and expert standards—with 114 failing the integrity check.
Return to original text
A short introduction is retained here; the full text from the source website is not reposted. For specific wording and subsequent updates, please refer to the original article.
Open AIHOT original