Guardian

World ‘may not have time’ to prepare for AI safety risks, says leading researcher

NDTV

AI Advancing Too Fast? Top Expert Warns World May Run Out of Time

MINT

We will be outcompeted: Leading AI expert warns world may be running out of time to prepare for AI risks

AI Security Institute

Frontier AI Trends Report by The AI Security Institute (AISI)

Advanced Research and Invention Agency

AI progress and a Safeguarded AI pivot

A look back at 2025

The Wall Street Journal

The AI Revolution Will Bring Prosperity

Metaculus

What % of current workers will be replaced by AI systems performing end-to-end labor in 2030?

New_AI_Safety

Elon Musk

Rishi Sunak

Kamala Harris

Bill Gates

Andrew Ng

Yoshua Bengio

Fei-Fei Li

Demis Hassabis

Melanie Mitchell

Sam Altman

Dario Amodei

Geoffrey Hinton

Yann LeCun

Gary Marcus

Max Tegmark

Connor Leahy

Eliezer Yudkowsky

Jaan Tallinn

Marc Andreessen

Eric Schmidt

Norbert Wiener

Arthur Clarke

Irving John Good

Claude Shannon

Hans Moravec

John Smart

There's a "wall of fear-mongering and doomerism" in the AI world right now.

"Worrying about AI today is like worrying about overpopulation on Mars."

Concerns that AI could pose a threat to humanity is "preposterously ridiculous."

Claiming AI poses an existential threat is "such an extreme" and risks "wip[ing] out some of its potential benefits."

"I'm more concerned about... the risks that are here and now [than the existential threat of AI]."

Current AI is "not anywhere close" to posing an existential threat but it could in the future.

AI is the "biggest existential threat" to humanity.

AGI's worst-case scenario would be "lights-out for all of us."

Powerful AI systems "taking control" pose an "existential threat."

AI has a "10 to 25 per cent" chance of destroying humanity.

"The most likely result of building a superhumanly smart AI... is that literally everyone on Earth will die."

Does AI Pose an Existential Threat to Humanity?

Joe Biden

Stuart Russell

Tristan Harris

Cédric O

"AI may cause the most dramatic and sustained economic boom of all time."

AI will make the world "way wealthier" and lead to a "productivity boom."

AI is an "amplifier of human intelligence" and will help the economy thrive.

AI will enable "universal high income" in place of "universal basic income."

"AI may become the heart of future economic growth in the coming decades."

AI harms could include "disrupting economies."

AI could "disrupt job markets."

"AI and technological unemployment" are the "biggest problem we face" from an "economic point of view."

"The moment superintelligent AI emerges, the human economy will end."

Will AI Bring Economic Risks?

Vladimir Putin

Hod Lipson

Yuval Noah Harari

Sundar Pichai

"[T]here will be far greater jobs" thanks to AI and "the jobs of today will get better."

AI will probably "take over a lot of jobs," but could also "create new kinds of jobs"

"[W]e might witness the creation of a massive new unworking class."

AI "is not going to put a lot of people out of work permanently" but work will change.

AI might take away more than just "drudge work."

"AI software will be in direct competition with a lot of people for jobs."

"Automation and AI will take away pretty much all of our jobs."

Eventually, "AI will be able to do everything" and jobs will exist only for "personal satisfaction."

Will AI Replace All Jobs?

David Dalrymple, a program director and AI safety expert at the U.K. government's Advanced Research and Invention Agency (ARIA), has warned that the world "may not have time" to prepare for the safety risks posed by AI.

David Dalrymple, a program director and AI safety expert at the U.K. government's Advanced Research and Invention Agency (ARIA), has warned that the world "may not have time" to prepare for the safety risks posed by AI.

Speaking to The Guardian, Dalrymple claimed it was "not science fiction" to imagine that within five years the "most economically valuable tasks" could be performed by machines "at a higher level of quality and lower cost than by humans."

Speaking to The Guardian, Dalrymple claimed it was "not science fiction" to imagine that within five years the "most economically valuable tasks" could be performed by machines "at a higher level of quality and lower cost than by humans."

The AI safety expert also advised that governments should not assume that advanced AI systems are reliable, noting that "the science to do that is just not likely to materialize in time given the economic pressure."

The AI safety expert also advised that governments should not assume that advanced AI systems are reliable, noting that "the science to do that is just not likely to materialize in time given the economic pressure."

The danger, according to Dalrymple, is the risk of creating advanced AI systems capable of outcompeting people "in all of the domains that we need to be dominant in, in order to maintain control of our civilization, society and planet."

The danger, according to Dalrymple, is the risk of creating advanced AI systems capable of outcompeting people "in all of the domains that we need to be dominant in, in order to maintain control of our civilization, society and planet."

Dalrymple's interview comes as the U.K. AI Security Institute published a report finding that AI capabilities were "rapidly improving" and doubling every eight months in some areas, with advanced AIs completing apprentice-level tasks 50% of the time on average, compared to over 10% in early 2024.

Dalrymple's interview comes as the U.K. AI Security Institute published a report finding that AI capabilities were "rapidly improving" and doubling every eight months in some areas, with advanced AIs completing apprentice-level tasks 50% of the time on average, compared to over 10% in early 2024.

In response to the accelerated growth of AI, Dalrymple's Safeguard program at ARIA pivoted in November from funding specialist AI teams to turning its TA1 toolkit into "a foundational, open-source platform for mathematical assurance, formal verification and auditable AI systems."

In response to the accelerated growth of AI, Dalrymple's Safeguard program at ARIA pivoted in November from funding specialist AI teams to turning its TA1 toolkit into "a foundational, open-source platform for mathematical assurance, formal verification and auditable AI systems."

Guardian #3444

NDTV #

MINT #

AI Security Institute #

Advanced Research and Invention Agency #

The Wall Street Journal #

Metaculus #

David Dalrymple, a program director and AI safety expert at the UK government's Advanced Research and Invention Agency, warned that the world may not have time to prepare for safety risks posed by rapidly advancing AI systems.

David Dalrymple, a program director and AI safety expert at the UK government's Advanced Research and Invention Agency, warned that the world may not have time to prepare for safety risks posed by rapidly advancing AI systems.

Dalrymple stated that it is not science fiction to project that within five years most economically valuable tasks will be performed by machines at a higher level of quality and lower cost than by humans.

Dalrymple stated that it is not science fiction to project that within five years most economically valuable tasks will be performed by machines at a higher level of quality and lower cost than by humans.

The UK AI Security Institute reported that AI capabilities are advancing rapidly, with performance doubling every eight months in some areas. Advanced models can now complete apprentice-level tasks 50% of the time, compared to just over 10% in early 2024.

The UK AI Security Institute reported that AI capabilities are advancing rapidly, with performance doubling every eight months in some areas. Advanced models can now complete apprentice-level tasks 50% of the time, compared to just over 10% in early 2024.

AISI tested models for self-replication and found two cutting-edge models achieved success rates over 60%. However, AISI stressed that a worst-case scenario is unlikely in everyday conditions.

AISI tested models for self-replication and found two cutting-edge models achieved success rates over 60%. However, AISI stressed that a worst-case scenario is unlikely in everyday conditions.

Dalrymple advised that governments should not assume advanced AI systems are reliable, noting that the science needed to ensure full reliability may not emerge quickly enough due to economic pressures.

Dalrymple advised that governments should not assume advanced AI systems are reliable, noting that the science needed to ensure full reliability may not emerge quickly enough due to economic pressures.

The Safeguarded AI programme pivoted in response to faster-than-expected frontier AI progress, with the program now focusing on expanding a toolkit for mathematical assurance and formal verification rather than developing specialized AI capabilities.

The Safeguarded AI programme pivoted in response to faster-than-expected frontier AI progress, with the program now focusing on expanding a toolkit for mathematical assurance and formal verification rather than developing specialized AI capabilities.

History proves technological revolutions create far more jobs than they destroy while dramatically raising living standards. From the Industrial Revolution doubling real wages to the Digital Revolution creating 5.1 million jobs monthly, markets consistently absorb disruption and deliver prosperity when government stays out of the way.

History proves technological revolutions create far more jobs than they destroy while dramatically raising living standards. From the Industrial Revolution doubling real wages to the Digital Revolution creating 5.1 million jobs monthly, markets consistently absorb disruption and deliver prosperity when government stays out of the way.

Advanced AI systems could outcompete humans across all critical domains within five years, yet civilization sleepwalks into this transition without adequate safety measures. Cutting-edge models already achieve 60% success rates on self-replication tests and complete expert-level tasks autonomously, moving faster than safety science can keep pace.

Advanced AI systems could outcompete humans across all critical domains within five years, yet civilization sleepwalks into this transition without adequate safety measures. Cutting-edge models already achieve 60% success rates on self-replication tests and complete expert-level tasks autonomously, moving faster than safety science can keep pace.

Artificial Intelligence

UK Expert Warns World May Lack Time for AI Safety Prep

Headlines

There is a possibility that advanced AI systems could outcompete people across all critical domains within the next five years. Yet, society is sleepwalking into this transition without any adequate safety measures. Governments must wake up to the danger, as well as the reality that the AI systems in which they have placed their faith are not as reliable as they hope.

There is a possibility that advanced AI systems could outcompete people across all critical domains within the next five years. Yet, society is sleepwalking into this transition without any adequate safety measures. Governments must wake up to the danger, as well as the reality that the AI systems in which they have placed their faith are not as reliable as they hope.

The coming AI revolution will bring with it unimagined prosperity, just like its historical counterparts, but only if the government stays out of the way. If society, at the behest of public outcry and fear, decides instead to restrict the growth of AI through regulation, not only will this delay the inevitable, but it will also make the world and its people poorer as a result.

The coming AI revolution will bring with it unimagined prosperity, just like its historical counterparts, but only if the government stays out of the way. If society, at the behest of public outcry and fear, decides instead to restrict the growth of AI through regulation, not only will this delay the inevitable, but it will also make the world and its people poorer as a result.

19.2% of current workers will be replaced by AI systems performing end-to-end labor in 2030, according to the Metaculus prediction community.

19.2% of current workers will be replaced by AI systems performing end-to-end labor in 2030, according to the Metaculus prediction community.

UK Expert Warns World May Lack Time to Prepare for AI Safety Risks

UK Expert Warns World May Lack Time to Prepare for AI Safety PrepRisks

The Spin

Metaculus Prediction

The Controversies

Andreessen

Ng

LeCun

Mitchell

Li

Bengio

Musk

Altman

Hinton

Amodei

Yudkowsky

Andreessen

Altman

LeCun

Musk

Bengio

Mitchell

Harris

Russell

Tallinn

Altman

Mitchell

Harari

LeCun

Hinton

Ng

Lipson

Musk

Go Deeper

Articles on this story

Sign Up for Our Free Newsletters
Sign Up for Our Free Newsletters

UK Expert Warns World May Lack Time to Prepare for AI Safety Risks

UK Expert Warns World May Lack Time to Prepare for AI Safety PrepRisks

The Spin

Metaculus Prediction

The Controversies

Andreessen

Ng

LeCun

Mitchell

Li

Bengio

Musk

Altman

Hinton

Amodei

Yudkowsky

Andreessen

Altman

LeCun

Musk

Bengio

Mitchell

Harris

Russell

Tallinn

Altman

Mitchell

Harari

LeCun

Hinton

Ng

Lipson

Musk

Go Deeper

Articles on this story

Sign Up for Our Free NewslettersSign Up for Our Free Newsletters

Sign Up for Our Free Newsletters
Sign Up for Our Free Newsletters