Improved AI safety, security sought by lawmaker

Google, Microsoft, Amazon, OpenAI, Meta, Inflection AI, and Anthropic have been urged by Senate Select Committee on Technology Chairman Mark Warner, D-Va., to further prioritize artificial intelligence security and safety that are not adequately addressed by the voluntary commitments they have given to the Biden administration, reports The Record, a news site by cybersecurity firm Recorded Future. While the companies' commitments to perform security testing, bolster cybersecurity investments, and clarify AI-generated content are admirable, Warner said the firms should not only ensure the implementation of industry standards across all their offerings, but also promise to curb AI use in the generation of nonconsensual intimate images, real-time facial recognition, and social-scoring efforts. "These commitments have the potential to shape developer norms and best practices associated with leading-edge AI models. At the same time, less capable models not covered by the commitments are susceptible to misuse, security compromise, and proliferation risks," said Warner, who also criticized Apple, Midjourney, and other AI firms for not entering the commitments.