Unverified50% confidenceFactExact time
In the study, AI handled only 23.3% of review comments, and the vast majority of inspection work was still done by humans.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Related Entities
Related Claims
UnverifiedEven the best-performing models can only recover 30% of the evaluated evidence and 25% of the limitations descriptions.65% similarUnverifiedA detection rate of 73.43% means that approximately one in four core errors still goes uncaught, indicating that automated review cannot yet fully replace the judgment of human experts.64% similarUnverifiedOnly 41.9% of AI Overviews were fully grounded, meaning every claim had support.60% similarUnverifiedWhen small models are made to do verbatim quotations, only 24% to 35% of the quotations pass verification against the source text.60% similarUnverifiedAmong claude-in-chrome calls, screenshot calls, which account for only 23%, contributed 49% of total token consumption.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1054960API
curl https://kongchang.com/api/v1/knowledge/claims/1054960MCP
get_claim(id=1054960)