szuperintelligencia (ASI)
in English: superintelligence (artificial superintelligence)
A hypothetical artificial intelligence that far surpasses even the best human experts in almost every intellectual task; it does not exist today.
In plain terms
Superintelligence is not a product but a vision. It refers to machine intelligence that is better than humans in almost everything, from scientific research and strategic planning to persuasion.
Imagine playing against a world chess champion. You know you will lose, but you cannot say which move will defeat you. Superintelligence would represent the same dominance, not on a chessboard, but across almost every field simultaneously.
The hope is that such a system could accelerate drug research, energy, or basic science. The fear is that a system smarter than us is much harder to control and keep aligned with our goals.
It differs from today's chatbots and agents in that those are already faster than you at some tasks but still make frequent mistakes and need human supervision. Superintelligence would mean orders of magnitude more than that.
How it works
- No one knows exactly how it would be built, because it does not yet exist; the following steps describe the most frequently cited concept.
- As a first step, systems would reach human-level capability in most intellectual work. This is called Artificial General Intelligence (AGI).
- Afterwards, artificial intelligence would also participate in its own development: agents would write the code, design experiments, and improve the next model generation.
- If this loop accelerates itself, capabilities could rise steeply in a short time. This is called an intelligence explosion, though it is debated whether it will actually occur.
- Physical limits also slow down progress: chips, electricity, data centers, and the time required for real experiments cannot be accelerated arbitrarily.
An everyday example
You read in the news that a major tech company is founding a 'superintelligence lab' and spending billions of dollars on data centers. In such cases, it is not about a ready product but a long-term goal that also serves as a message to investors. The assistant running on your phone remains the same regardless.
Who is affected
A common misconception
Limits and risks
- There is no unified, measurable definition, so anyone can claim to be close to it, and this is hard to refute or verify.
- The term is often used for marketing and fundraising purposes, so company leaders' predictions should be treated with caution.
- The question of control remains unresolved: we cannot be sure how to reliably align a system smarter than us with human goals.
- Debate about the distant future can distract from today's real AI problems, such as errors, data privacy, and workplace impacts.
Similar, but not the same
- artificial general intelligence (AGI) – AGI roughly means human-level capability in most intellectual tasks, while superintelligence far exceeds this.
- singularity – The singularity is a hypothetical turning point or period when development accelerates unpredictably, while superintelligence is the system itself.
- Narrow artificial intelligence – Narrow AI can be superhuman in a single area, such as chess or protein structure prediction, while superintelligence would be so in almost everything.
Related concepts
artificial intelligence, futures studies
speed superintelligence, collective superintelligence, quality superintelligence
general AI, intelligence explosion, recursive self-improvement, singularity, AI safety, alignment, existential risk
A brief history
British mathematician Irving John Good wrote about the "ultra-intelligent machine" and the intelligence explosion in 1965. The term superintelligence was widely popularized by philosopher Nick Bostrom's 2014 book, Superintelligence.
Where things stand
Superintelligence does not exist even at the end of 2026. Current systems are strong in specific tasks but do not surpass humans overall. Several major AI companies openly state it as a goal, while there is a sharp professional debate about the timing and feasibility.
In the news
- Democrats promise AI regulation as they prepare to take control of US CongressNov 4, 2026 · Jog
- Microsoft diagnostic AI outperforms generalist doctors on complex casesOct 22, 2026 · Agent Age
- Microsoft's MAI-DxO in Science: twice as accurate as doctors on hard casesOct 22, 2026 · Health
- GitHub outlines "ambient intelligence" direction with physical Copilot controlsOct 19, 2026 · Code
- Hinton, Bengio and 400 scientists call for international body to oversee advanced AIOct 13, 2026 · Agent Age
- Trump administration has leading tech firms sign voluntary AI safety accordSep 28, 2026 · Agent Age
- Meta unveils Muse agent and Charm device as Amazon blocks itSep 21, 2026 · Agent Age
- Sanders and Casar bill would ban development of superintelligenceSep 21, 2026 · Agent Age
Also known as
szuperintelligencia · artificial superintelligence · szuperintelligens MI · superhuman artificial intelligence · ASI · superintelligence · artificial superintelligence · superhuman AI
Sources
This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.