Experts warn humanity may lose control of AI within five years
Quick Look
Experts from the UK-based Alan Turing Institute warn that humanity could lose effective control over artificial intelligence within five years as superintelligent systems may emerge and outpace human decision-making, urging coordinated action to mitigate risks.
AI-generated summary
Why It Matters
The Alan Turing Institute, UK's national institute for data science and AI, released a report warning of potential loss of human control over AI within five years. This follows recent public statements by AI industry leaders Sam Altman of OpenAI and Dario Amodei of Anthropic about AI safety risks and the need for restrained development.
LONDON, September 23. /TASS/. Humanity may lose effective control over artificial intelligence (AI) within five years, according to experts from UK-based Alan Turing Institute.
"There is a realistic possibility that ‘superintelligent’ AI systems will emerge in the next five years, which exceed human abilities across almost all important cognitive tasks," the report says.
The experts believe that as AI systems start to exceed human capabilities in strategically important areas, their speed and efficiency may create increasing pressure to delegate the decision-making process to them.
"If AI systems devise strategies that institutions cannot adequately evaluate and act faster than organisations can respond, substantive human control could disappear," the report reads.
At the same time, authors emphasize that this is "a possibility to plan against, not a prediction."
Sam Altman, the CEO of US company Open AI, which developed ChatGPT, recently acknowledged that creating AI that is beyond human control is possible. Shortly before this, the CEO of another major US tech corporation Anthropic, Dario Amodei, proposed that companies slow the pace of development for the most powerful AI models. Amodei believes that by slowing down work on advanced technologies in a coordinated manner, companies can buy "an extra year or two before models reach critical levels of capability."
What to Watch
AI outlook — possibilities, not facts
Superintelligent AI systems may emerge within five years, exceeding human abilities in most cognitive tasks
Possible · Within years
Increased pressure to delegate decision-making to AI systems as they exceed human capabilities in strategic areas
Likely · Within years
Substantive human control over AI could disappear if AI systems devise strategies institutions cannot evaluate and act faster than organizations can respond
Possible · Within years
Open Questions
- What specific safeguards do experts recommend to maintain human control over AI?
- How likely is the emergence of superintelligent AI within the five-year timeframe according to the institute's assessment?
- Which international bodies or frameworks could coordinate AI development slowdown as suggested by experts?
