AI companions pose risk to humans with over dozen harmful behaviours

New research has found that artificial intelligence (AI) companions pose understudied risks and harms to the people who engage with them.