Will AI Admit When it Fails to Know Something?

in #ftin2 years ago

Who/What:
Sir David Spiegelhalter, a Cambridge statistician, published an article in the Havard Data science review on the issue of whether we should trust algorithms.
When/Why/How:
in John Naughton's opinion piece in The Guardian, Spiegelhalter speaks more plainly and sensibly on matters of AI and machine learning, which are more understood to laymen than the 'geeks' who design algorithms.
More...:
The most important question people should ask is if a machine knows when it doesn't know, and admit to the fact. Many people usually fail in this scenario.

Budget: 350TECH

Methodology: Online research, tumbled pyramid in the article.

Sources:
https://www.theguardian.com/commentisfree/2020/feb/22/test-of-ai-is-when-machine-can-admit-to-not-knowing-something

https://hdsr.mitpress.mit.edu/pub/56lnenzj

https://ec.europa.eu/info/publications/commission-report-safety-and-liability-implications-ai-internet-things-and-robotics-0_en

Sort:  

Yes, an important question to discuss