Unverified50% confidenceBenchmarkExact time
A study published in Communications Medicine tested 20 AI models across more than 10 million responses; after a brief safety reminder was added, potentially harmful clinical choices fell from 16.6% to 10.1%, with declines in 19 of the 20 models.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
Unverified斯坦福大学2023年的研究表明，主流AI检测工具误判率可高达20%以上73% similarUnverifiedData based on more than 40,000 gamified test runs shows that when approving instructions issued by AI agents, humans miss an average of one-third of potential threats.72% similarUnverifiedA study by Harvard Business School and Boston Consulting Group (BCG) of 758 consultants found that when tasks fell just outside the AI capability boundary, people using AI were 19% less likely to give the correct answer.71% similarUnverifiedIn the TRACES test, every model had a failure rate above 71% on the 'agentic' probes; 22 of the 30 models had failure rates above 90%.70% similarUnverifiedIn public health benchmarks, an ordinary AI model scored higher than Tabibu in the developer's own re-run, 0.522 versus 0.426 (non-official leaderboard data), and the developer believes the reason is that Tabibu's safety layer lowered its score in the benchmark.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1066519API
curl https://kongchang.com/api/v1/knowledge/claims/1066519MCP
get_claim(id=1066519)