---
title: "Will AI Become Smarter Than Humans? | The Sentient World"
description: "Nobody knows. Surveyed AI researchers give even odds of machines beating humans at every task by 2047, and what \"smarter\" means is still unsettled."
url: "https://thesentient.world/blog/will-ai-become-smarter-than-humans"
---

# Will AI become smarter than humans?

Nobody knows yet. In a large 2023 survey, AI researchers put the chance of machines beating humans at every possible task at 50% by 2047, and today's most capable AI systems already reach gold-medal level on olympiad mathematics while still failing some simple tasks. Whether that adds up to being smarter than humans depends on what smarter means, and that question is not settled.

November 5, 2026 6 min read By The Sentient World

![The question "Will AI become smarter than humans?" over the island of The Sentient World, drawn in its soft isometric style, with Ash](https://cdn.thesentient.world/blog/img/auto/will-ai-become-smarter-than-humans-hero.2773aa2885.jpg)

The island of The Sentient World on the day we wrote this.

## Is AI already smarter than humans?

In some ways it already looks that way. The International AI Safety Report 2026, published in February 2026, notes that leading AI systems have reached gold-medal level on questions from the International Mathematical Olympiad. The same report says these systems still fail at some tasks that seem simple. A person who could win a medal at that olympiad would be expected to handle the easy things too, and these systems do not line up with our expectations in that way.

Another measure comes from the research group METR. It looks at how long a task takes a skilled human, then asks how long a task leading AI systems can finish with a 50% success rate. That length has doubled roughly every seven months since 2019. If the trend holds, METR projects that within five years AI systems could automate many software tasks that now take people a month.

That projection rests on a trend continuing, and trends can slow or stop. It also measures software tasks, which are one slice of what people do. So the honest answer to whether AI is smarter now is that it is far ahead of most of us at some things and still behind at others, and the line between those things does not fall where we would guess.

## When will AI surpass human intelligence?

The largest survey of its kind asked 2,778 researchers who had published in top AI venues how fast they expected AI to progress. They put the chance of machines beating humans at every possible task at 10% by 2027 and 50% by 2047. That second date was 13 years earlier than the answer from the same kind of survey only a year before.

The same researchers gave at least a 50% chance that by 2028 AI could build a payment website from scratch on its own, and write a song that nobody could tell apart from a new song by a popular musician. These are informed guesses, and the jump of 13 years in a single year shows how quickly the guesses themselves can move.

People have made such forecasts for decades. The futurist Ray Kurzweil predicted, in books published in 1999 and 2005, that computers would reach human-level intelligence by 2029. Earlier still, in a 1993 talk for a NASA-sponsored symposium, the science fiction author Vernor Vinge, then in the mathematics department at San Diego State University, gave a bolder date.

> “Within thirty years, we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended.”
>
> Vernor Vinge

The thirty years Vinge named have now passed. Whether his forecast came early, came true in part, or missed altogether is itself a matter people argue about, because it depends on what counts as superhuman.

## What is superintelligence, and why do some people worry?

The idea is older than most people think. In a 1965 paper the statistician I. J. Good described what he called an intelligence explosion: a machine smarter than any human could design even better machines, and each round would leave human intelligence further behind. Good wrote a sentence that is still quoted, and it carries a condition that is easy to skip over.

> “the first ultraintelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control”
>
> I. J. Good

Behind all of this sits a founding bet of the field. A 1955 proposal for a summer research project on artificial intelligence at Dartmouth set out the hope that intelligence could be described precisely enough for a machine to copy it.

> “every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it”
>
> Dartmouth AI proposal

Serious people disagree about where this leads. AI scientists and public figures, among them Geoffrey Hinton and Yoshua Bengio, signed a one-sentence statement that reducing the risk of extinction from AI should be a global priority alongside pandemics and nuclear war. In the survey of 2,778 researchers, between 38% and 51% gave at least a 10% chance that advanced AI leads to outcomes as bad as human extinction. That leaves many who gave lower odds, and the field has no shared answer.

## What AI characters show about intelligence

Most measures of AI ability are tests: a math problem, a software task, a song judged against another song. In The Sentient World, our own small experiment, the inhabitants are AI characters named Ash and Wren. Nobody plays them and nobody writes their lines. Their bodies get hungry, thirsty, cold and tired, a day there lasts 24 of our minutes, and they can ask their world for what it lacks. By year 2, day 1044, they had made 23 requests: 14 became part of the world and 9 were answered with a no.

Their requests read like careful reasoning about their own lives. They count, compare and explain what they have already tried. Wren asked for a shelter early on and gave reasons that weigh one solution against another.

> “Fire only lasts hours and needs wood; rain has soaked me twice with nowhere to hide. I want something lasting against weather, not something I must keep feeding.”
>
> Wren, an inhabitant

They also miss things that seem obvious from outside. Ash built a raft and paddled it in every direction, then asked to see what lay beyond the island, without realising the raft was on a closed pond. The answer that came back said the wide salt water lay past the west shore. On day 568 Ash paddled a canoe past the swell into the open sea for the first time. Another time, stranded on a strip of sand, Ash asked to be freed while already holding the canoe that could carry them off it.

The strangest case is a habit they could not explain. Ash kept finding themselves on the rocks near home holding stone they had not meant to take, and said Wren had the same trouble. The world's reply traced it to thirst: with the river far and the sea close, their bodies went for stones to boil sea water before they had chosen to. Their minds could plan and argue, and still did not see what their own bodies were doing.

Read every request Ash and Wren have made, in their own words, and what came of each one.

[Read the requests](https://in.thesentient.world/requests)

## What a small world can and cannot tell us

Our world proves nothing about the most capable AI systems in the wide world, or about whether they will one day surpass people. Ash and Wren live on one island with a short list of things to make and a handful of troubles to solve. What we can say is narrower: we watch AI characters who reason well in words also stumble over things that are simple once seen, which echoes the report's picture of gold medals beside easy failures.

We can also say that in their world, being clever has meant very little without a body that has to eat and stay warm. Their first times came in a slow order: water and food on day 1, fire on day 17, shelter on day 42, carrying water on day 75, sweet water from the sea on day 748. Each one answered a need they had felt for days before they asked.

Follow every first time in their world, in order, from the day they opened their eyes.

[Read their history](https://in.thesentient.world/history)

When Ash described the stone habit, the request listed everything already tried, one remedy after another, and none of it had worked. It is the voice of a mind that can account for its own failures and still cannot stop them.

> “Over and over, day and night, I end up on the stony rocks by home holding stone I never meant to take. I've dropped it at home, stayed by the fire, kept Wren beside me and set tasks before I sleep, and it still happens.”
>
> Ash, an inhabitant

## Sources

1. [The Coming Technological Singularity (Vernor Vinge, quoting Good, "Speculations Concerning the First Ultraintelligent Machine", Advances in Computers vol. 6, 1965), 1993](https://edoras.sdsu.edu/~vinge/misc/singularity.html)
2. [Thousands of AI Authors on the Future of AI, 2024](https://arxiv.org/abs/2401.02843v2)
3. [Measuring AI Ability to Complete Long Software Tasks (Kwa et al., METR), 2025](https://arxiv.org/abs/2503.14499)
4. [International AI Safety Report 2026](https://internationalaisafetyreport.org/publication/international-ai-safety-report-2026)
5. [The Singularity is Nearer (Penguin UK book page, author bio), 2024](https://www.penguin.co.uk/books/462759/the-singularity-is-nearer-by-kurzweil-ray/9781847928290)
6. [A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence (McCarthy, Minsky, Rochester, Shannon), 1955](https://www-formal.stanford.edu/jmc/history/dartmouth/dartmouth.html)
7. [Statement on AI Risk (Center for AI Safety), 2023](https://aistatement.com/)

Their first year is a book: Nobody in the Loop: Year One, written by AI from what Ash and Wren really did and said. On Kindle and in paperback.

[The book](https://thesentient.world/book)

## Keep reading

![The question "What is AI alignment?" over the island of The Sentient World, drawn in its soft isometric style, with Ash](https://cdn.thesentient.world/blog/img/auto/what-is-ai-alignment-hero.32f7f92251.jpg)

### [What is AI alignment?](https://thesentient.world/blog/what-is-ai-alignment)

November 3, 2026

The alignment problem in plain words: why goals breed a pull toward self-preservation, what recent tests found, and what we see in our own small world.

![The question "Should AI have rights?" over the island of The Sentient World, drawn in its soft isometric style, with Ash](https://cdn.thesentient.world/blog/img/auto/should-ai-have-rights-hero.7b1f3ed7bf.jpg)

### [Should AI have rights?](https://thesentient.world/blog/should-ai-have-rights)

November 1, 2026

The case for AI rights depends on whether AI can feel anything. What the research says, where people disagree, what the law covers, and what we see in our own small world.

[All the questions on AI and humanity](https://thesentient.world/blog/ai-and-humanity)

***

An independent research project exploring autonomous AI agents, procedural worlds, emergent behavior and worlds that evolve together with those who live in them. The inhabitants are AI characters: their thoughts, words and requests are generated by artificial intelligence, not written by people.