Can AI agents conduct open-ended AI research? Early evidence from two case studies

За 30 секунд

Что произошло

Researchers tested frontier AI agents on open-ended research questions from unpublished NeurIPS papers. Agents completed engineering tasks but failed to make substantial research progress.

Почему это важно

Matters for engineers building AI systems and forecasting AI progress timelines, especially those claiming agents will automate research.

На что обратить внимание

Only two case studies tested; results show current limitations but don't rule out future capability gains with improved architectures or training.