Superintelligence (ASI)

What it means

Superintelligence, or ASI (artificial superintelligence), is an AI that would beat the best people at nearly every kind of thinking, not just one. Picture your top strategist, your best scientist, and your most persuasive negotiator, then picture one system better than all three at each of their jobs. A chess program is better than any human at chess and nothing else. A superintelligence would have that kind of lead almost everywhere.

No one has built one. The people trying to build it describe it as a goal: a public statement hosted by the Future of Life Institute says that "many leading AI companies have the stated goal of building superintelligence in the coming decade."¹

Where the word comes from

The idea is older than the word's current use. In 1965 the statistician I. J. Good described an "ultraintelligent machine": "a machine that can far surpass all the intellectual activities of any man however clever." He concluded that "the first ultraintelligent machine is the last invention that man need ever make."²

In a paper first published in 1998, the philosopher Nick Bostrom gave the modern term a definition that people still cite: "By a 'superintelligence' we mean an intellect that is much smarter than the best human brains in practically every field."³ For many years the word belonged mostly to philosophers and futurists. Since then it has become a business goal. One company is named for it, Safe Superintelligence Inc., and says "Superintelligence is within reach."⁴ Meta has written that "Developing superintelligence is now in sight."⁵ In October 2026 it reached government, as the name of a White House task force.⁶

How others define it

• Nick Bostrom (1998): "an intellect that is much smarter than the best human brains in practically every field," "including scientific creativity, general wisdom and social skills."³
• I. J. Good (1965), on the "ultraintelligent machine": "a machine that can far surpass all the intellectual activities of any man however clever."²
• The Statement on Superintelligence: a system that "can significantly outperform all humans on essentially all cognitive tasks."¹
• Future of Life Institute press release: "artificial intelligence capable of outperforming all humans at most cognitive tasks."⁷

Where they agree and differ

They agree on the core idea: superintelligence goes beyond the best humans, across many fields at once. That is what sets it apart from [AGI](/definitions/agi), which is usually described as matching people, not passing them.

They differ on how wide the lead has to be. Bostrom says "practically every field," the statement says "essentially all cognitive tasks," and the institute's own press release says "most."³¹⁷ Bostrom also adds a line that matters to leaders: "Entities such as companies or the scientific community are not superintelligences according to this definition."³ A group of smart people working together does not count, however capable the group is.

The biggest split is over whether to build it. The statement calls for "a prohibition on the development of superintelligence, not lifted before there is" "broad scientific consensus that it will be done safely and controllably, and" "strong public buy-in."¹ The strongest view on the other side comes from the builders. Meta writes, "I am extremely optimistic that superintelligence will help humanity accelerate our pace of progress," and adds, "We believe the benefits of superintelligence should be shared with the world as broadly as possible."⁵ Even Meta grants that "superintelligence will raise novel safety concerns."⁵ Both sides take the risks seriously. They disagree about whether to slow down.

In the news

On October 4, 2026, President Donald Trump announced what he called a "Super Intelligence Force," which ABC News described as "an artificial intelligence task force."⁸ Its stated aim is to "ensure that America continues to lead the World in Super Intelligence," and Trump said Director of National Intelligence Jay Clayton would lead it.⁸ TechCrunch reported that the task force will "reportedly have 120 days to create a report on the risks and opportunities presented by AI."⁶ Neither report includes a definition of the term. The only description TechCrunch quotes is Trump's own: something "many say is bigger than the Industrial Revolution, and the Internet."⁶ We are confident the task force was announced. What it will actually do is not yet known.

What it means for your organization

The AI agents your organization uses today are not superintelligent. But the word now appears in company strategies, product pitches, and government policy, and people often use it without saying what they mean. When you see it, ask which definition the speaker has in mind: better than the best humans at everything, or at "most" things.

The practical lesson holds no matter how capable a system becomes. Someone has to decide what it may do, and a person has to stay accountable for it. This publication calls that rule [a person in the middle](/definitions/a-person-in-the-middle).

Question for your team: When a vendor says its product is a step toward superintelligence, what can the product actually do for us today, and which person approves what it does?