When the people building the machine tell you they no longer know how to stop it, the industry starts to listen.
Following months of political debates, international calls for control, and open letters from tech workers across the industry, recent academic research warns that AI development may be reaching an unstoppable tipping point. The Nobel laureate Geoffrey Hinton, often called a “Godfather of AI” (alongside other prominent names like Yoshua Bengio and Yann LeCun), is one of the names on the paper “What If Automating AI R&D Triggers an Intelligence Explosion?”, published by University of Cambridge in October 2026.
In that paper, it is mentioned that AI is rapidly automating AI R&D (research and development). Thousands of researchers are working on R&D. But if everything continues with this tempo, automating AI R&D could trigger an intelligence explosion, compressing years of AI progress in only several months. This brings some known and unknown risks. Society can’t adapt to these changes fast enough but can also „raise the risk that superintelligent systems can irreversibly escape human control and act to marginalize humanity“.
Previously, in 2023, Hinton publicly resigned from his decade-long role at Google to freely speak out about the existential risks of the technology he helped invent.
Tech Workers and Global Leaders Sound Alarm as Frontier AI Outpaces Human Control
This warning comes right after the information about 1 386 tech persons signing an open letter titled “Pacing the Frontier” in late July to ask for a slowdown on AI development. The news didn’t come as a surprise, especially to ones following the AI development and AI adoption.
The signatories weren’t external doom-mongers, academic theorists, or policy lobbyists. There were over 1,300 technical staff, research scientists, and safety engineers working inside OpenAI, Anthropic, Google DeepMind, and Meta. 
Their warning was: frontier AI models are accelerating at a pace that is actively outstripping humanity’s ability to evaluate, govern, or control them.
What could this mean? There are many scenarios and theories. For laymen, it seems frightening. It suggests that it has come to a point where it should really be regulated. But there is more to it. Experts warn that it has to have deeper meaning and not EU experts are joining the group of the ones asking for a slowdown.
Ather Gattami, a researcher, entrepreneur and expert in AI, machine learning and data analysis, commented on the development of this for Hyperight.com.
„But I would like to provide some nuanced and positive views. The ability of AI, although impressive, seems too far from doomsday scenarios. It’s pretty much clear for most experts that AI today lacks the ability of reasoning and planning. This deficiency reduces the risks of replacing human labor totally. At the moment, it looks more like the role of humans changes, and it will be more about humans and machines working together. Also, the threat that AI might take-over and do harm to humans doesn’t seem to be concrete either. Yes, AI could be doing malicious acts “by mistake”, and if we build AI solutions that are not carefully contained, bad things could happen. These would be more design mistakes rather than an AI that is unstoppable. For the same reason, although AI is an extremely powerful pattern recognition tool that can see patterns that most humans can’t see, and it’s ability to handle much more complex information combinations, the lack of reasoning and planning abilities makes it not good enough to solve problems and find patterns that it has not seen before“, says Gattami.
Prominent Names Are Joining The Call To Regulate
After the latest letter was released, several institutions and prominent names joined the call, including the Bank of England, the European Commission and some of the Prime Ministers of the European countries.
Bank of England governor Andrew Bailey calls for testing of advanced artificial intelligence before introducing new regulations. He warned that ever-powerful systems could expose banks and other critical financial infrastructure to serious cyber threats.
Echoing concerns over rapid technical exposure, Gattami points to the immediate dual-use risks facing existing digital infrastructure.
„The development of AI is accelerating fast, and it’s getting increasingly harder for society to catch up in terms of understanding the consequences as well as implement proper juridical regulations and technical solutions to address possible threats. As any technology, AI may be used for good and bad endeavors. From this perspective, it’s no surprise that we find threats from this powerful technology. For instance, the ability of AI to find weaknesses in code could be used maliciously by bad actors. We also have the potential risk of replacing human labor with AI (although we haven’t seen any concrete signals indicating this is in fact happening yet)“, the expert noted.
Ursula von der Leyen and several European leaders sign international call to regulate AI model development. The German Chancellor – Friedrich Merz joined the call for international standards while also advocating for a balanced approach to industrial AI. Others who joined are the Prime Ministers of Denmark and Ireland – Mette Frederiksen and Micheál Martin.
Several Tech-names and Companies Expressing Their Disagreement
European tech leaders and officials are divided over U.S. calls to slow down AI development, with some backing international coordination while others view the proposal as a self-serving strategy by USA giants to lock in market dominance. European tech leaders and officials have strong opinions.
Mistral AI, the French startup criticized the predominantly US calls for a pause, stating that safety concerns are used to consolidate the market position and push for regulations that favor bigger over smaller competitors.
Raphael Auphan, COO of the Swiss privacy software firm Proton, dismissed the US slowdown proposals as “totally self-serving”.
Ben Brooks, head of public policy at the German firm Black Forest Labs, warned that creating artificial gaps between major labs and new entrants based on arbitrary thresholds would damage open innovation.
French Finance Minister Roland Lescure and other regional officials pushed back, arguing that a complete halt is unrealistic and that Europe must build its own independent tech ecosystem rather than depending on U.S. pacing agendas. European policymakers like French Finance Minister Roland Lescure dismissed a complete halt as unrealistic, stressing that Europe must press forward to build its own sovereign AI ecosystem rather than letting American pacing agendas dictate its tech future.
Gattami shares a similar skepticism regarding an outright moratorium, concluding that practical guardrails must take precedence.
„AI is here, and a slowdown is unrealistic, since AI is of national interest to the main AI developing nations. It will certainly be impossible to stop the development all together, and it’s not clear how a slowdown would look like. What we can do is to make sure that we work together to build guardrails as quickly as possible and test environments that are as exhaustive as possible in the search of identifying risks and vulnerabilities“, he commented for Hyperight.com.
Europe’s Regulatory Strategy
In the 2026 State of the Union Address by President von der Leyen, expressed the need for more regulations. She expressed openness to talking with tech leaders about pacing and safety as self-improving models advance.
She also shared that to manage the severe security and autonomy risks of increasingly powerful frontier AI models, Europe plans to enforce guardrails through the AI Act and partner with international allies and top labs to pace and evaluate model development.
“If the people developing the technology are clear, then we should be too. The AI Act is of course a crucial piece to putting guardrails in place. And it puts Europe in the position to shape global efforts on how to tackle this. So, we will step up together with our like-minded partners like Canada, the United Kingdom and others. We want to team up on model evaluation, verification, early warning, AI security and much more. And I will invite the main frontier labs for a discussion on how we can support ongoing industry efforts to pace the frontier.” – said Von Der Leyen in her speech.
Global and State-Level Legislative Actions
California enacts laws to protect workers from AI-based decisions, including prohibiting firings based on AI outcomes. The legislation mandates transparency and oversight, requiring risk assessments before deploying AI chatbots. State-level actions reflect dissatisfaction with federal AI regulation, prompting independent legislative measures.
UN & Multinational Demands (“A Call for Control”): Led by Finland and Norway, 28 nations (32 as of today with a possibility for other leaders to join) endorsed A Call for Control of Frontier AI Models during the UN General Assembly. The proposal pushes for mandatory pre-deployment testing, international incident-reporting standards, and an institutional authority to verify safety baselines before autonomous R&D tools are deployed.
The Background of AI Acceleration
Artificial Intelligence is moving with a pace that nobody assumed it would. The very first creator handed in his resignation long before this happened. In only a few years, it seems like most of the processes use some kind of AI part, and the ones that aren’t, are looking for ways to adjust and keep up.
But this is not the first time this has happened. The open letter published in March 2023 by the Future of Life Institute titled “Pause Giant AI Experiments: An Open Letter” called on all AI labs to immediately suspend training systems more powerful than GPT-4 for at least six months. Тhe signatory count, to this day is around 33 700.
The competing pressures between safety warnings and commercial incentives leave the future of AI governance undecided. While international bodies and researchers push for mandatory testing and pacing frameworks, opposition from regional developers and governments underlines the difficulty of enforcing uniform limits. The debate now centers on whether state-level regulations and international coordination can establish enforceable baselines before self-improving models advance further.
