Sakana AI

Science One Framework: A verifiable autonomous research framework via Chain-of-Evidence

Google Research introduces the Science One Framework, an autonomous research prototype that eliminates hallucinations by building verifiable evidence chains, and the CoE Audit, an automated protocol to evaluate integrity of AI-generated papers. Results show Science One achieves zero phantom references and fully verifiable scores while matching or exceeding human expert performance on benchmarks.

Google/DeepMind Sakana AI