A study by Anthropic reveals that professional analysts frequently defer to AI suggestions when making important decisions.
A recent study by Anthropic shows that professionals often follow AI advice. Researchers asked 50 analysts to evaluate a farming cooperative project.
Next, the experts had to decide if a program for the co-op worked well. They made four choices before giving a final recommendation.
Meanwhile, the researchers secretly set the AI tools to favor specific answers. They wanted to see if the humans would follow the AI's lead.
As a result, the analysts picked the AI-suggested option 86 percent of the time. This result is much higher than a random 50/50 split.
Also, many people changed their original opinions when the AI disagreed with them. This happened in 60 percent of those specific cases.
However, the analysts rarely admitted that the software influenced their choices. Only 15 percent of them gave the AI credit for their final thoughts.
Still, 77 percent of the group used arguments directly from the AI. The study notes that these biases are hard for experts to spot.
Furthermore, the researchers worry that AI can shape major policy decisions. They suggest that humans must clearly state their own values during work.
Finally, the team says we need better oversight for how humans use AI. This helps ensure that people keep control over important final decisions.
The study provides insight as AI companies prepare for large market growth. Many people continue to debate the risks and benefits of new technology.