AI’s performance is limited by the quality of the data it has been trained on—it can only be as good as its training data, never exceeding it. This limitation arises because AI excels at interpolation (making predictions within the range of its training data) but struggles with extrapolation (generalizing to entirely new or unknown situations). Additionally, AI lacks true logical reasoning abilities and cannot draw conclusions that extend beyond the information it has been explicitly trained on. In contrast, humans have the ability to reason logically, connect seemingly unrelated ideas, and apply abstract thinking to solve problems in entirely new contexts. Humans can extrapolate creatively, draw insights from incomplete data, and apply intuition to situations they have never encountered before—qualities that remain beyond the capabilities of AI. This makes humans far better at adapting to novel challenges and generating original ideas.
Published by Anders
Anders Hast received a PhD in Computerised Image processing at Uppsala university in 2004. In 2011 he spent one year at IIT, CNR, Pisa in Italy as an ERCIM fellow and after that he received a full time position as associate professor and in 2019 he became professor in image processing, both at Uppsala University. He has been affiliated with UPPMAX super computer centre where he worked as an application expert in Scientific Visualisation and later in the field of digital humanities together with the Centre for digital humanities in Uppsala. His recent research has focused on image processing and computer vision, for applications in microscopy, aerial photography, object recognition and especially hand written text recognition and face recognition. View all posts by Anders
Anders Hast 