Daijiworld Media Network – Washington

Washington, Nov 12: Microsoft’s AI CEO Mustafa Suleyman has revealed that the tech giant was previously barred from developing its own Artificial General Intelligence (AGI) models under an agreement with OpenAI — a restriction that remained in place until recently.

In an interview with Business Insider, Suleyman said the original deal prevented Microsoft from pursuing AGI independently until 2030, forcing it to focus only on smaller AI efforts and fine-tuning OpenAI’s models. “Microsoft needs to be self-sufficient in AI,” he said.

Following a renegotiation, Microsoft is now free to build its own AGI, leading to the creation of a new division — Microsoft AI Superintelligence — dedicated to frontier AI research. The team will work on challenges such as transfer learning and continual learning to develop human-like adaptive systems.

This marks a shift from partnership to competition, positioning Microsoft against AI leaders like OpenAI, Google, Meta, Anthropic, and xAI. Despite the rivalry, Suleyman said Microsoft remains open to using models from other firms, including GPT and Claude, depending on user needs.

Microsoft is also expanding its AI infrastructure, investing in custom chips and massive data clusters, while prioritising safety under new Responsible AI VP Trevor Callaghan.

“There’s a risk that these systems get extremely smart and run away from us,” Suleyman cautioned. “We must ensure a humanist intent that keeps humans at the top.”

With this newfound freedom, Microsoft is preparing for a new phase of AI innovation — one where it competes at the very frontier of intelligence.