On Thursday the CEO of Microsoft AI Mustafa Suleyman announced in a blog post the formation of a new team to perform advanced AI research. The team, called the MAI Superintelligence Team, will be led by Suleyman and will focus on instilling a humanist philosophy in AI development. “The history of humanism has been its enduring ability to fight off orthodoxy, totalitarian tendencies, pessimism and help us preserve human dignity.” Suleyman continued “In that spirit, we think this approach will help humanity unlock almost all the benefits of AI, while avoiding the most extreme risks.”
Towards A Humanist AI Approach
Suleyman has been in the AI business for over a decade, co-founding AI startup DeepMind in 2010 before that company was purchased by Google in 2014. He left Google in 2022 to found the AI startup Inflection, before he was hired last year to lead Microsoft AI, which includes Bing and the AI assistant Copilot. Suleyman’s announcement of a team to help ensure that the development of AI follows a human-aligned and philosophically deliberate path comes at a time of skepticism for the industry. Concerns about runaway costs and potentially catastrophic misaligned AI have grown as investments in AI development take up an increasingly significant share of domestic economic growth.
While AGI (Artificial General Intelligence) is commonly understood to be the point at which AI models can perform as well as expert humans in any domain, superintelligence would then be the point at which AI goes well beyond human capabilities. This is why the question of alignment, of creating a “Humanist Superintelligence” (HSI) as Suleyman describes, weighs so heavily on those in the AI research field.
The MAI Superintelligence Team intends to help address both issues of costs and the threat of misalignment in the AI industry broadly. “There is at present a collective action problem of more unsafe models of superintelligence potentially being able to develop faster and operate more freely.” Suleyman said, adding in the same post “That’s why I’m publishing this – to start a process and to make clear that we are not building a superintelligence at any cost, with no limits.”
The Potential of Aligned Superintelligence
While the worries over AI development are serious, the announcement of Microsoft’s superintelligence team is not without a heavy dose of optimism. “AI is the path to better healthcare for everyone. AI is how our society levels up, escapes an increasingly zero-sum world. It’s how we grow the economy to increase wealth broadly, and enable a higher standard of living across society.” Three main areas of focus for the MAI Superintelligence Team will be personal assistants, advanced healthcare, and energy abundance, where Suleyman suggests HSI will allow humanity to make rapid breakthroughs. “They will make a transformative difference to billions of people. This next decade may well be the most productive in history.”
It is not uncommon for the major AI developers to state their intentions and expectations as such, and to attempt to distinguish their approach to alignment as more sincere and sophisticated than their competitors. It is a good reminder of the stakes of the AI project to hear such a clear vision of alignment from an industry leader like Suleyman. But the question of whether Microsoft’s MAI Superintelligence Team will live up to such lofty aims, or any of the major AI players for that matter, still remains to be answered.
