Encountered a problematic response from an AI model? More standards and tests are needed, say researchers

AI has a security problem — and industry insiders say testing standards have a long way to go

from US Top News and Analysis https://ift.tt/o9EeGmp