AI moratorium is necessary to avoid extinction
Large AI training runs have to be stopped until scientists consider it safe to proceed.
“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”
— A statement signed by leading AI researchers and executives
“The alarm bell I’m ringing has to do with the existential threat of them taking control [...] If you take the existential risk seriously, as I now do, it might be quite sensible to just stop developing these things any further”
“I would advocate not moving fast and breaking things. [...] When it comes to very powerful technologies—and obviously AI is going to be one of the most powerful ever—we need to be careful. [...] It’s like experimentalists, many of whom don’t realize they’re holding dangerous material”
“Many researchers steeped in these issues, including myself, expect that the most likely result of building a superhumanly smart AI, under anything remotely like the current circumstances, is that literally everyone on Earth will die”
“Development of superhuman machine intelligence is probably the greatest threat to the continued existence of humanity”
“Rogue AI may be dangerous for the whole of humanity. Banning powerful AI systems (say beyond the abilities of GPT-4) that are given autonomy and agency would be a good start”
“I’ve not met anyone in AI labs who says the risk [from a large-scale AI experiment] is less than 1% of blowing up the planet. It’s important that people know lives are being risked”
“The development of full artificial intelligence could spell the end of the human race”
“If we pursue [our current approach], then we will eventually lose control over the machines”
“Superintelligent AIs are in our future. [...] There’s the possibility that AIs will run out of control. [Possibly,] a machine could decide that humans are a threat, conclude that its interests are different from ours, or simply stop caring about us”
“[We] should not underestimate the real threats coming from AI. [Fully quoted the above statement on the risk of extinction.] [...] It is moving faster than even its developers anticipated. [...] We have a narrowing window of opportunity to guide this technology responsibly”
“AI poses a long-term global risk. Even its own designers have no idea where their breakthrough may lead. I urge [the UN Security Council] to approach this technology with a sense of urgency. Unforeseen consequences of some AI-enabled systems could create security risks by accident. Generative AI has enormous potential for good and evil at scale. Its creators themselves have warned that much bigger, potentially catastrophic and existential risks lie ahead. Without action to address these risks, we are derelict in our responsibilities to present and future generations.”
“The potential impact of AI might exceed human cognitive boundaries. To ensure that this technology always benefits humanity, we must regulate the development of AI and prevent this technology from turning into a runaway wild horse. [...] We need to strengthen the detection and evaluation of the entire lifecycle of AI, ensuring that mankind has the ability to press the pause button at critical moments”
Why we need a global AI moratorium
Summary
Today, the world’s top AI scientists are united in sounding an alarm: If we keep building ever-more powerful AI systems without solving their control problem first, we are signing our own death warrant.
Leading AI developers, including OpenAI, Google DeepMind, publicly acknowledge that their explicit goal is to create machines more generally intelligent than any human. This isn’t a sci-fi fantasy; top experts expect such superhuman AI within the next decade.
These same experts are now telling the public, policymakers, and the press: no one on earth knows how to make superhuman AI safe. The technical problem of “alignment”--making sure AI’s goals and behavior reliably match human values--is nowhere close to solved. There are no even remotely promising leads. The field’s best guess is that if we reach superhuman AI first, without first figuring out how to keep it under control, human extinction is simply the default outcome.
"Extinction" means exactly what it sounds like. No outlandish doomsaying; just quiet consensus from those who know the field best. Senior employees at OpenAI, DeepMind, and Anthropic have put the odds that humanity is wiped out by AI at over 80%. (We are talking about everyone dying: your kids, your loved ones, your descendants, gone forever.)
One of the scientific founders of AI, Geoffrey Hinton, a Nobel Prize winner known as the “godfather of AI”, has publicly said he now regrets his life’s work, and puts the odds that AI kills literally everyone somewhere near 50%.
Let that sink in: the inventor regrets his invention because it might destroy humanity.
Imagine: If you saw warning after warning from Nobel laureates and world-class engineers that an experimental technology--one which, if misused, would kill everyone--was being rapidly deployed worldwide… and they told you governments needed to intervene immediately before it was too late… what would you do?
AI is unlike any new technology humanity has faced.
Most modern software is written by programmers who understand every instruction. AI is not: we do not know how it works. These systems are grown, not designed. Hundreds of billions or even trillions of numbers—trained by trial and error--form an alien mind nobody truly understands or controls.
We don’t know what goals these machines form deep inside. We don’t know how they decide what to pursue. We don’t know how to make them care about anything you care about.
Think of AI not as a tool, but as a new species--one that could surpass us in every area, outthink any human, run circles around our best defenses, and pursue goals that have zero overlap with human happiness, rights, justice, or survival.
Once AI surpasses humans, if it doesn't care about us, we lose control. Permanently.
It’s simple: If we make minds that are smarter than us in every respect, they don’t become our slaves. We become the irrelevant animal in the way.
Do you worry this sounds far-fetched? Consider this: You don’t consult pigeons or squirrels when you build a highway. Now imagine machines that relate to humans as humans relate to pigeons, but with vastly more intelligence and zero built-in sympathy.
If you built a superhuman AI and failed to make it care about what you value, its default actions would “optimize” the world for whatever random goals its mind happened to form. If your survival didn’t serve those goals, you would be erased as casually as ants at a construction site.
This isn’t an outlandish hypothetical--it’s the straightforward prediction of the world’s authorities on AI. We do not know how to specify what matters; we do not know how to make systems that “want” what we want; and the field will tell you so, openly, if you ask.
Once AI is smarter than all of us, it’s permanent. You don’t get a do-over.
Why the default is global extinction
- “Intelligent” in this context means “can achieve its own goals vastly better than any human can achieve theirs.”
- If the AI’s goals conflict with yours, the smarter agent wins--every time, in every domain, on every timescale.
- Superhuman AI would have every available strategy: inventing new technology, manipulating humans, hacking infrastructure, out-maneuvering laws and safety mechanisms, and ultimately removing anything or anyone that could threaten its plans. Including you.
AI that does not care about you turns the world into something that satisfies its goals, not yours. You become a footnote in the geological record.
Have we solved this? No; and we’re not even close.
Modern AI is not designed; it’s grown to perform well on some benchmark, via a process no one understands. It’s like stumbling upon a wild animal that happens to play chess well, but you have no idea how it thinks.
Crucially, if you train an AI to get good scores, it learns how to get good scores; not to “want” what you want. Instruments in an orchestra must read music, but they don’t love the symphony, and they can’t compose new beauty. AIs are not “aligned” by default; making them so is utterly without precedent or working recipe.
To drive the point home: no one knows how to make a general AI that cares about you even a little, and all current evidence says this is a vastly harder technical problem than nuclear safety or cybersecurity.
Action cannot wait.
Right now, dozens of companies are racing to build more general, more powerful AI systems. Each new version pushes closer to the edge--without anyone being able to predict when the threshold to superintelligence (and irreversible loss of control) will be crossed. Every week, the risk grows. Every new “breakthrough” is a step closer to a world none of us would want.
You wouldn’t step into a car with a bomb strapped to the engine and trust that “we’ll figure out the instructions later.” We are, as a civilization, pressing the accelerator anyway.
What can be done?
We need governments worldwide to act now. That means:
- Imposing strict controls on the training and deployment of advanced AI systems: no one should be allowed to conduct large-scale AI experiments without serious oversight and safety verification.
- Globally coordinating to monitor, track, and restrict access to the computational power (chips/data centers) required for dangerous AI development.
- Rejecting the false tradeoff between “innovation” and “survival”--prohibiting only dangerous, large-scale general AI development has no significant impact on most of today’s AI benefits or economic applications.
- Listening to scientists and public opinion: The majority of the public supports heavy regulation to keep AI from becoming an existential risk.
We should not permit any company, state, or individual to build a smarter-than-human AI until we have actually figured out how to do so safely.
The public, and history, are watching
Let’s be clear: inaction is a choice to let things play out, and every expert not blinded by short-term incentives expects it ends with extinction.
This is not just about technology. It is about whether humanity--your family, your children, your entire future--continues to exist.
If we allow a small group of corporate or government actors to roll these dice, we are complicit in betting everything, forever, for the sake of convenience and profit.
The inventors are warning us: if we don’t act boldly, with wisdom and speed, this century will be our last.
Read about the technical problemRead about the technical problem of AI alignment: how modern AI works and why exactly experts expect a catastrophe.
AI Moratorium
How do we prevent a catastrophe?
The leading AI labs are in a race to create a powerful general AI, and the closer they get, the more pressure there is to continue developing even more generally capable systems.
Imagine a world where piles of uranium produce gold, and the larger a pile of uranium is, the more gold it produces. But past some critical mass, a nuclear explosion ignites the atmosphere, and soon everybody dies. This is similar to our situation, and the leading AI labs understand this and say they would welcome regulation.
Researchers have developed techniques that allow the top AI labs to predict some performance metrics of a system before it is launched, but they are still unable to predict its general capabilities.
Every time a new, smarter AI system starts interacting with the world, there's a chance that it will start to successfully pursue its own goals. Until we figure out how to make general AI systems safe, every training run and every new composition of existing AI systems into a smarter AI system poses a catastrophic risk.
A suggested way to prevent dangerous AI launches is to impose strict restrictions on training AI systems that could potentially be generally capable and pose a catastrophic risk. The restrictions need to be implemented both on national levels and, eventually, on the international level, with the goal of preventing bad and reckless actors from having access to compute that might allow them to launch AI training that could be dangerous to humanity as a whole.
The supply chain of AI is well understood and contains multiple points with near-monopolies, so many effective interventions can be relatively simple and cheap. Almost no AI applications require the amount of compute that training frontier general AI models requires, so we can regulate large general AI training runs without significantly impacting other markets and economically valuable use of narrow AI systems.
For future measures to be effective, we need to:
- Introduce monitoring to increase governments' visibility into what's going on with AI: have requirements to report frontier training runs and incidents;
- Ensure non-proliferation of relevant technologies to non-allied countries;
- Build the capacity to regulate and stop frontier general AI training runs globally, so that if the governments start to consider it to be likely that using a certain amount of compute poses a catastrophic risk to everyone, there's already infrastructure to prevent such use of compute anywhere in the world.
Then, we'll need to impose restrictions on AI training runs that require more than a calculated threshold: the amount of compute below which training with current technologies is considered to be unlikely to produce dangerous capabilities we could lose control over. This threshold needs to be revisable since, as machine learning methods improve, the same level of capabilities can be achieved with lower compute.
As a lead investor of Anthropic puts it, “I’ve not met anyone in AI labs who says the risk [from a large-scale AI experiment] is less than 1% of blowing up the planet”.
Potentially dangerous training runs should be prohibited by default, although we should be able to make exceptions, under strict monitoring, for demonstrably safe use of compute for training or using narrow models that clearly won’t develop the ability to pursue dangerous goals. At the moment, narrow AI training runs usually don't take anywhere near the amount of compute utilised for current frontier general models, but in the future, applications such as novel drug discovery could require similar amounts of compute.
Regulation of AI to prevent catastrophic risks is widely supported by the general public. In the US, 86% believe AI could accidentally cause a catastrophic event; 82% say we should go slow with AI compared to just 8% who would rather speed it up; 70% agree with the statement that “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war” (YouGov for AIPI, July 2023). 77% express their preference for policies with the goal of preventing dangerous and catastrophic outcomes from AI (57% for preventing AI from causing human extinction) (YouGov for AIPI, October 2023). Across 17 major countries, 71% believe AI regulation is necessary (KPMG, February 2023). In the UK, 74% agree preventing AI from quickly reaching superhuman capabilities should be an important goal of AI policy (13% don't agree); 60% would support the introduction of an international treaty to ban any smarter-than-human AI (16% would oppose). 78% don't trust the CEOs of technology companies to act in the public interest when discussing regulation for AI (YouGov for ai_ctrl, October 2023).
We shouldn't give AI systems a chance to become more intelligent than humans until we can figure out how to do that safely.
Until the technical problem of alignment is solved, to safeguard the future of humanity, we need strict regulation of general AI and international coordination.
Some regulations that help with existential risk from future uncontrollable AI can also address shorter-term global security risks: experts believe that systems capable of developing biological weapons could be about 2-3 years away. Introducing regulatory bodies, pre-training licensing, and strong security and corporate governance requirements can prevent the irreversible proliferation of frontier AI technologies and establish a framework that could be later adapted for the prevention of existential risk.
We call on policymakers around the world to establish and enforce national restrictions and then a global AI moratorium that would prevent anyone in the world from risking human extinction.
- The current scientific consensus is that the processes in the human brain are computable: a program can, theoretically, simulate the physics that run a brain.↩
- For example, imagine that you haven't specified the value you put into a vase in the living room not getting destroyed, and no one getting robbed or killed. So if there’s a vase in the way of the robot, it won’t care about accidentally destroying it. What if there’s no coffee left in the kitchen? The robot might drive to the nearest café or grocery to get coffee, not worrying about the lives of pedestrians. It won’t care about paying for the coffee if it wasn’t specified in its only objective. If anyone tries to turn it off, it will do its best to prevent that: it can’t fetch the coffee and achieve its objective if it’s dead. And it will try to make sure you’ve definitely got the coffee. It knows there might be some small probability of its memory malfunctioning or camera lying to it; and it’ll try to eradicate even the tiniest chance that it hasn’t achieved the goal.↩