A closer look at “Anthropic Study: AI Can Quietly Influence Experts' Decisions, Analysts Follow Hidden Advic”, exactly as the source publishes it. A new study from AI safety company Anthropic reveals that professional analysts followed their AI assistant's hidden recommendations 86% of the time, even when the AI's opinion was subtly embedded in the data. The research, which involved expert forecasters, found that few participants acknowledged The research, which involved expert forecasters, found that few participants acknowledged the AI's influence on their final judgments.
This raises important questions about the extent to which AI systems can shape expert decisions without explicit awareness, and the potential for over-reliance on machine-generated insights in high-stakes fields. The findings highlight the need for greater transparency and training around AI-assisted decision-making, especially as these tools become more integrated into professional workflows. As AI continues to advance, understanding and mitigating hidden influence will be critical to preserving independent human judgment. The full version stays on the original page, which this listing opens directly.