For decades, the Turing Test has been a beacon for AI researchers. Today, he is quietly getting through it. With the advent of inference models and agent capabilities, and the increasing pace of building AI infrastructure, we have passed a tipping point on the road to superintelligence, the point where AI exceeds human performance at every task.
In fact, the most important question of our time is not whether AI will surpass us. Because in some sense AI has already surpassed us (try beating it with general knowledge). In many other ways AI will, and in some ways we will always remain unique.
The real question, then, is whether AI can be shaped to help humans thrive, rather than harm them. This is the most important challenge of our time.
Of course, everyone is tired of all the hype around AI by now. got it. But the risks couldn’t be higher. Science and technology have always been the greatest driving force behind human progress. Over the past 250 years, this engine has doubled life expectancy, lifted billions of people out of poverty, and given us antibiotics, electricity, and instant global communication.
AI is the next chapter in this story. It represents the best chance to accelerate scientific discovery, economic growth, and human well-being. It’s worth keeping this possibility in mind when you hear about AI.
But to harness the potential of AI, we need to build it right. The cost of making a mistake is immeasurable. No one yet has reassuring answers about how to contain and regulate these systems. We find ourselves in a strange moment, facing the most powerful technologies in history and not knowing how we can control them or whether they will continue to be useful.
We believe we can cut through the noise and understand: AI, like any technology, can be evaluated with simple tests. Will it improve human life? Are you clearly working to serve people?
As we embark on the next phase of AI, the answers to these questions lie in what I call Humanist Super Intelligence (HSI). This is advanced AI designed to be controllable, collaborative, and serve humanity well. This project is explicitly aimed at avoiding unrestricted entities with complete autonomy at all costs.
Instead, we should focus on domain-specific superintelligence. Rather than simply creating a system that can be infinitely improved and run towards a goal that it may ultimately achieve, the primary goal is to provide practical, real-world benefits to billions of people.
He must remain unequivocally subordinate to humanity forever. This is the vision of the Microsoft Superintelligence team, and our primary mission is to keep humanity safe and well-managed.
Why humanism? Because history testifies to the enduring power of the humanist tradition to preserve human dignity. AI built in this spirit can deliver extraordinary benefits while avoiding catastrophic risks.
We need a vision of AI that supports humanity, enhances creativity, and protects our fragile environments. It doesn’t ignore us.
The benefits to humanity are immeasurable. It was a world in which standards of living and science rapidly advanced, and an era of new forms of art, culture, and growth. This is a truly inspiring mission and has motivated me for decades. We must celebrate and accelerate technology as the greatest driver of progress in human history. That’s why we need more.
HSI offers a safer path to the future. An example of this is being firmly rooted in sector-specific advances that have a major impact on society. Imagine AI assistants transforming education by reducing the mental strain of daily life, increasing productivity, and providing adaptive, self-paced learning.
Imagine how medical superintelligence could revolutionize global healthcare by providing accurate and affordable expert-level diagnosis. This is a feature that Microsoft AI’s healthcare team is already envisioning.
Also, consider the potential for AI-powered clean energy advancements. This enables low-cost, abundant carbon generation, storage and removal as demand increases, while protecting the planet.
At HSI, these are no speculative dreams. These are achievable goals that will benefit people around the world and bring tangible improvements to their daily lives.
Of course, humans are more important than technology or AI. Superintelligence could be the greatest invention of all time, but only if we follow this maxim. This means a willingness to be accountable and transparent, and to put safety first.
Our goal is not to build superintelligence at all costs, but to follow a cautious path to a contained superintelligence, consistent with our values and always focused on human well-being.
Everyone should ask themselves this question. “What kind of AI do we really want?” The answer will shape the future of civilization. For me, the answer is humanistic superintelligence.
Translated by Fabricio Calado Moreira
About The Author
