The safeguards include a mix of preexisting policies and newer initiatives to prevent the misuse of its wildly popular generative AI tools.