Microsoft has formally entered the race toward artificial superintelligence (ASI) with a new initiative dubbed “Humanist Superintelligence,” led by AI chief Mustafa Suleyman. In an interview with Axios and a company blog post, Suleyman emphasized that the project aims to build a powerful AI that serves humanity rather than merely maximizing performance. He stated, “The project of superintelligence has to be about designing an AI which is subservient to humans, and one that keeps humans at the top of the food chain.” (axios.com)
This announcement follows Microsoft’s renegotiated deal with OpenAI and reflects the company’s intent to catch up in the increasingly competitive and costly race toward AGI and ASI. Suleyman’s framing of the initiative as “humanist” underscores a growing industry emphasis on safety and alignment, even as regulatory attention shifts elsewhere. (axios.com)
The broader context reveals a surge in ASI-related discourse and investment. SoftBank’s Masayoshi Son has predicted that ASI—systems vastly more intelligent than humans—could emerge within a decade, potentially becoming 10,000 times smarter than human brains. He envisions ASI contributing up to 10% of global GDP within 10–15 years, rising to 30% over 30 years. (time.com)
Meanwhile, academic research continues to explore ASI’s theoretical underpinnings and safety frameworks. One study proposes a “multi-box protocol” for aligning ASI via mutual verification among isolated systems, aiming to ensure truthful behavior through peer validation. (arxiv.org) Another research effort redefines “superalignment” as a human–AI co-alignment toward a sustainable, symbiotic society, combining external oversight with intrinsic AI empathy and self-awareness. (arxiv.org)
Taken together, these developments mark a pivotal moment in the ASI narrative. Industry leaders are not only accelerating toward superintelligent systems but are also foregrounding alignment and human-centric design. As Microsoft’s initiative illustrates, the next phase of AI development may hinge as much on values and governance as on raw capability.