John Ratcliffe, the C.I.A. director, said that the agency would take “smart risks,” but that people would have oversight of artificial intelligence.