How do you stop people from blindly accepting AI answers?
Design it like this
For high-stakes decisions, ask for the user's own take first, then reveal the AI suggestion.
What the research says
Small 'speed bumps', like asking people to decide before revealing the AI's answer, reduced over-reliance. The trade-off: people liked these designs less.
“However, there was a trade-off: people assigned the least favorable subjective ratings to the designs that reduced the overreliance the most.”
Buçinca et al., 2021
Sources
- To Trust or to Think: Cognitive Forcing Functions Can Reduce Overreliance on AI in AI-assisted Decision-making2021 · CSCW 2021 · Zana Buçinca, Maja Barbara Malaya, Krzysztof Z. Gajos
“However, there was a trade-off: people assigned the least favorable subjective ratings to the designs that reduced the overreliance the most.”
Abstract