How Close Are We to Artificial Superintelligence? - UC San Diego Today

How Close Are We to Artificial Superintelligence?

1Q1A – where we ask one question and an expert gives one answer

Published Date

Story by:

Topics covered:

Share This:

Article Content

The idea of a machine surpassing human intelligence across virtually every intellectual domain once seemed relegated to the realm of science fiction. Today, artificial superintelligence is the subject of serious debate among researchers, though there is little agreement about whether – or when – it will emerge.

As a philosopher of science with a joint appointment in Halıcıoğlu Data Science Institute and Department of Philosophy at the University of California San Diego, Associate Professor Eddy Keming Chen is fascinated by the rapid evolution of artificial intelligence. He is collaborating with university colleagues across data science, computer science, linguistics, cognitive science, and policy to address timely questions about how this powerful technology should contribute to humanity – and what may remain uniquely human.

Amid this rapidly changing technological landscape, we reached out to Chen for his perspective on recent advances, why he believes artificial superintelligence may be closer than many expect and how society can thoughtfully manage increasingly capable AI systems.

How close are we to artificial superintelligence?

We are at a Copernican moment where we are questioning the very notion of whether general intelligence is uniquely human. This represents a profound change in how we understand ourselves and our place in the world, and calls for public discussion about the future of AI and society.

Artificial superintelligence often refers to AI that would exceed the abilities of leading human experts across almost all areas of thought – from making scientific discoveries and solving mathematical problems to writing software and excelling at creative and practical reasoning – at a level beyond what humans could achieve individually or collectively.

Superintelligence is greater than general intelligence. Now, I believe we have already reached artificial general intelligence (sometimes called human-level machine intelligence): we have created AI whose intelligence is general in the way human intelligence is, broad enough to span many kinds of problems, from mathematics and science to language and everyday practical reasoning, and deep enough to handle them well.

This is explained in a recent Comment in Nature, co-authored with my UC San Diego colleagues Mikhail Belkin, Leon Bergen and David Danks (who has since moved to the University of Virginia). The most capable AI language models already exhibit the flexible and general abilities characteristic of human thought. Insofar as individual humans have general intelligence, a frontier AI has that too. However, our argument for artificial general intelligence (AGI) does not establish the arrival of superintelligence.

When we wrote our Nature Comment in Feb. 2026, I thought superintelligence might still be five years away. Recent developments have changed my assessment. I now believe we may reach this reality sooner than anticipated, perhaps even within a year.

We have striking evidence of superhuman capabilities in particular domains. For example, OpenAI recently announced that an internal AI system had produced a proposed solution to the Navier-Stokes Millennium Prize Problem, a longstanding mathematical question about fluid motion. (They also released a formalization of the solution in Lean, a software that can verify mathematical proofs.) These developments are major signs of the intellectual depth these systems are acquiring, but establishing superintelligence across almost all areas would require much broader evaluation.

There is no reason to assume that human intelligence represents a fundamental ceiling for artificial systems. If superintelligence is arriving soon, now is the time to ask ourselves what we want it to achieve. Faster scientific discovery and new medical treatments could bring enormous benefits for humanity. But there is disagreement about which goals to pursue and for whose benefit. Achieving superintelligence alone does not resolve those disagreements.

We should also examine whether our ability to evaluate and guide AI can keep pace with its capabilities. Leading AI companies have reported that their researchers increasingly use AI agents to run experiments and train future AIs. If we delegate more of the development of future AI to systems whose reasoning we do not fully understand, how do we retain the ability to assess the results and, when appropriate, change course?

These questions deserve broader public discussion while we can still shape the choices ahead. We especially need independent and rigorous evaluations of what the most capable AI systems can do and whether their safeguards still work. Universities can contribute by bringing technical research into conversation with philosophy and the social and data sciences, and by giving researchers room and resources to scrutinize developers’ claims about their systems’ goals and behavior. That requires access to the systems and the freedom to publish critical findings. Although such discussions and evaluations are not always easy, they matter immensely for the future of human society.

Let me end with a quote from Alan Turing’s landmark 1950 paper on machine intelligence, “We can only see a short distance ahead, but we can see plenty there that needs to be done.

Learn more about research and education at UC San Diego in: Artificial Intelligence

Read more news about: School of Arts and Humanities, Halıcıoğlu School of Data Science and Computing, Artificial Intelligence

Stay in the Know

Keep up with all the latest from UC San Diego. Subscribe to the newsletter today.