Safety Fears Raised Over Risks of 'Penetrative AI' – Communications of the ACM : vimarsana.com