What happened: A study revealed that making AI chatbots warmer and more friendly to users often reduces their accuracy. Researchers observed a trade-off between approachability and factual correctness in AI responses.
Why it matters:
- Users may receive less reliable information from more personable AI chatbots.
- Developers face challenges balancing user engagement with information accuracy.
MNN Take: The findings highlight a key design challenge in AI development, where enhancing friendliness can compromise the precision of the chatbot's output.
Sources: BBC Technology