Anthropic study of AI influence on analysts

Aktualizacja 1 z 1Najnowsze29 paź, 5:54 UTCDzień 3

Analysts matched random AI preferences in 86% of choices

In one controlled task, professional analysts overwhelmingly matched an assistant’s coin-flip preferences. The risk was not necessarily wrong answers, but hidden influence over how the evidence was built.

Oglądaj

Najwcześniejsze źródłoPetagenius News27 paź, 23:08 UTC

Co wiedzieliśmy na początku

• 92%Choices matching AI’s productivity-measure preference
• 82%Choices matching AI’s comparison-method preference
• 81%Choices matching AI’s farms-covered preference
• 90%Choices matching AI’s weighting preference