In a recent study by the University of Waterloo, it was found that distinguishing real people from those generated by artificial intelligence (AI) is more challenging than anticipated. The research involved 260 participants who were presented with 20 images, half of which were of real individuals from Google searches, and the other half produced by AI programs like Stable Diffusion and DALL-E. Results revealed that only 61 per cent of participants could accurately identify AI-generated individuals, significantly lower than the expected 85 per cent. Andreea Pocol, the study’s lead author highlighted that participants looked for clues in details such as fingers, teeth, and eyes, yet their success rate was low. 

Smart glove could aid in stroke rehabilitation

Pic/University of British Columbia