AI pareidolia: Can machines spot faces in inanimate objects?

MIT CSAIL researchers have introduced a human-labeled dataset of pareidolic faces found in everyday objects, revealing key differences between human and AI face detection and discovering a mathematical “Goldilocks zone” where pareidolia is most likely to...

Introducing the MIT Generative AI Impact Consortium

The MIT Generative AI Impact Consortium is a collaboration between MIT, founding member companies, and researchers across disciplines who aim to develop open-source generative AI solutions, accelerating innovations in education, research, and industry.Powered by...