People are less likely to believe an AI if it conveys uncertainty
When a large language model expresses doubt about the information it supplies, people are less likely to accept it as fact and more likely to find accurate information elsewhere
![People are less likely to believe an AI if it conveys uncertainty](https://images.newscientist.com/wp-content/uploads/2024/06/04113355/SEI_207318483.jpg)
What's Your Reaction?
![like](https://news.sciencex.in/assets/img/reactions/like.png)
![dislike](https://news.sciencex.in/assets/img/reactions/dislike.png)
![love](https://news.sciencex.in/assets/img/reactions/love.png)
![funny](https://news.sciencex.in/assets/img/reactions/funny.png)
![angry](https://news.sciencex.in/assets/img/reactions/angry.png)
![sad](https://news.sciencex.in/assets/img/reactions/sad.png)
![wow](https://news.sciencex.in/assets/img/reactions/wow.png)