Microsoft Backs Independent AI Safety Evaluators and Middle East Expansion

Microsoft Corp. President Brad Smith announced the company's support for independent evaluators to ensure artificial intelligence safety during the United Nations General Assembly in New York. Smith argued that safety requires multiple layers of oversight rather than relying solely on leading labs such as OpenAI and Anthropic PBC. He advocated for the implementation of off-switches to maintain human control over advanced systems, suggesting these controls be managed by both governments and companies.

Beyond safety policy, Smith confirmed Microsoft will invest an additional $2 billion in the Middle East. This expansion includes utilizing 200 megawatts of computing capacity at a data center in the United Arab Emirates through a partnership with G42.

Smith also acknowledged that the company must adjust its 2030 carbon-negative climate goals. He noted that the rapid deployment of natural gas-fired plants, necessary to power AI data centers, has forced a reconsideration of the strategies Microsoft established six years ago.