Archive · 23 Sep 2026
UK Institute Warns Humans Could Lose Control of Superintelligent AI
A British research institute published a warning about very powerful artificial intelligence.
It said machines might become better than people at almost every important thinking task within five years.
Humans could begin relying on these machines for important decisions.
If the machines act faster than people and use plans humans cannot understand, people might lose meaningful control.
The report also warns that advanced AI could create chemical or biological dangers.
It could make false information spread and harm democratic systems.
Some AI researchers have called for development to pause or become more carefully controlled.
The report recommends practical safety research, strong rules and cooperation between countries.
This is a warning about a possible future, not a claim that superintelligent AI already exists.
The Alan Turing Institute says superintelligent AI could realistically emerge within five years.
The report warns humans may gradually hand critical decisions to AI systems that improve themselves.
Human control could disappear if AI develops strategies institutions cannot evaluate or respond to quickly.
The institute highlights potential chemical, biological, informational and democratic risks from advanced AI.
AI researchers and industry figures, including Dario Amodei, have urged stronger safeguards or a pause in frontier AI development.
- Who
- The Alan Turing Institute, along with AI researchers and industry figures concerned about advanced AI.
- What
- A report warns that superintelligent AI could emerge within five years and potentially exceed meaningful human control.
- When
- The report was published on Wednesday, September 23; it discusses risks that could arise within the next five years.
- Where
- The warning came from the UK-based Alan Turing Institute.
- Why
- The institute says increasingly capable AI could make decisions faster than institutions can assess them, while creating chemical, biological, informational and democratic risks.
- Reporting institution
- The Alan Turing Institute
- Report title
- Frontier AI poses credible near-term risks — demanding practical research, robust regulation, whole-system verification and international cooperation
- Estimated timeline
- A realistic possibility of superintelligent AI emerging within five years
- Central concern
- Human authority could disappear if AI systems devise strategies institutions cannot adequately evaluate
- Additional risks
- Chemical and biological threats, damage to the information environment, and harm to democratic systems
- Proposed safeguards
- Practical research, robust regulation, whole-system verification and international cooperation
- Industry response
- Dario Amodei proposed third-party evaluators, common safety benchmarks and international agreements, alongside pausing frontier AI development