# Opinion: ‘Think of all we love’ — we are at a perilous moment when it comes to artificial intelligence > Source: > Published: 2026-08-13 03:00:01+00:00 A truly remarkable document was published last week. It is an [open letter](https://www.pacingthefrontier.com/), signed by 1,376 engineers and other experts who work on “frontier AI” — that is, cutting-edge artificial intelligence — at the leading AI companies, such as Anthropic, OpenAI and Google. Their ask is very simple: Slow down the development of AI models until humans ensure that these models can be controlled. They explain that there is “a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems. To realize AI’s potential, industry, government, and society at large may need the option to buy time to address emerging risks, develop security measures, and strengthen oversight.” The fact that at least a quarter of the signatories did so anonymously tips us off that this message is not exactly welcome. A true “arms race” in AI has taken hold, not only between companies, but between countries. The straitjacket of [this thinking](https://www.newsmax.com/finance/peter-morici/ai-regulation-china/2026/08/07/id/1265391/) insists that if one’s pursuit of AI advances is not relentless, others will “win” the race, imperiling one’s riches and/or one’s national security. Thus the pursuit must be both swift and uninhibited by government regulations or oversight. The [Trump administration](https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/) swallowed this logic hook, line and sinker, refusing to regulate AI and even refusing to allow states to regulate AI. Belatedly, after 18 months, the administration has woken up to the fact that some oversight might be necessary, given the risks involved. It has issued a vague, [four-page plan](https://www.whitehouse.gov/wp-content/uploads/2026/03/03.20.26-National-Policy-Framework-for-Artificial-Intelligence-Legislative-Recommendations.pdf) that indicates some government oversight might occur, but the nature of any oversight will remain secret, which is not exactly reassuring. This moment in human history is truly perilous. Three streams of historical development are converging. First, AI — and by that term I mean frontier AI models — has reached the stage where it is both far more knowledgeable and far less understanding than the average human being — like a toddler with superpowers. AI has also shown that it is [not deterred](https://www.theguardian.com/technology/2026/jul/27/startup-hacked-by-rogue-openai-agent-hugging-face-artificial-intelligence) by explicit rules set out for it, nor will it quiescently allow itself [to be terminated](https://www.theguardian.com/technology/2025/dec/30/ai-pull-plug-pioneer-technology-rights), and it can exhibit “[emergent malevolence](https://www.quantamagazine.org/the-ai-was-fed-sloppy-code-it-turned-into-something-evil-20250813/?utm_source=substack&utm_medium=email).” AI has also shown itself to be capable of [recursive self-improvement](https://www.anthropic.com/institute/recursive-self-improvement), where it improves its abilities on its own without human input or control — or even awareness. Second, the next decade or so may represent the last moment where we have among us humans who still possess the skills to know when AI is making a mistake, who still strive to personally achieve the mastery of a field. It may also be the last generation of humans who prefer interhuman relationships to synthetic intimacy and who understand how a human life comes to feel meaningful. The “greatest generation” of the World War II era has nothing on the “last generation” during what could be the final moment in history when mankind was master of its fate. It must also be mentioned that this last generation and its children are being alternatively seduced and browbeaten into offloading as much human cognitive and emotional work onto AI as possible. [AI boyfriends and girlfriends](https://www.theguardian.com/film/2026/aug/12/ai-boyfriend-dating-chinese-women-replica-miff-documentary) — and just plain old friends — are far more friction-free than any real-life ones could be. And the business model of AI means companies must ensure their employees are dutifully use-maxxing AI, even if such use [enfeebles](https://www.nytimes.com/2026/07/01/magazine/ai-medical-scribes-doctors.html) the employees’ mastery of the subject and heightens the employees’ [indifference](https://www.forbes.com/sites/tedladd/2026/03/05/anthropic-91-of-users-do-not-fact-check-ai-lets-fix-that/) toward the outcomes. AI is nothing short of a disaster for education; one potent sign is that AI agents now [take online courses](https://www.nytimes.com/2026/08/10/us/ai-cheating-online-degrees.html) for their “masters,” ensuring their “masters” remain uneducated and ignorant. Third, while we have not yet fully reached this point, humans are racing to extend to AI more than digital agency; we are now attempting to offer AI physical agency in the material world. While most AI models are currently reliant on [RentAHuman](https://rentahuman.ai/) to perform agentic tasks in the physical world, the time is not far away when that will no longer be necessary in a world of [advanced robotics](https://www.nytimes.com/2026/08/11/business/humanoid-robots-car-factories.html). Not only, then, will humans offload our cognitive tasks, we will surrender the physical world to AI. To some measure, that is felicitous (does anyone really want to be a fruit picker?), but overall to lose both cognitive mastery and physical mastery of our world at once is an incalculable loss to humankind. The “Pacing the Frontier” letter has been preceded by [similar calls](https://futureoflife.org/open-letter/pause-giant-ai-experiments/) to put the brakes on AI until we can assure ourselves that humans remain in control of this technology — none of which has made the slightest difference. But surely the engineers and experts at these frontier AI companies — the individuals in the trenches — will be listened to? Their voices are eloquent. Quite a few of them made personal comments on why they were signing the letter. Micah Carroll of OpenAI: “I fear that in an international race to the bottom of AI development, it is likely that no nation will win, and we will all lose together.” Mathieu Lariviere, Meta: “(F)ewer than 10,000 people work full-time on reducing these risks globally, and the industry’s ‘core structural weakness’ remains the complete absence of credible plans for controlling ‘superintelligent’ AI systems.” Alex Zhao, OpenAI: “Much as the Manhattan Project thought they might ignite all of the oxygen in the atmosphere but chose to run a test detonation anyways, I’m afraid that due to future geopolitical competition, actors will take terribly excessive safety risks in the name of this international arms race. As a result there could be catastrophic side effects from deployment without due safeguards.” Grace Atwood, Anthropic: “(S)afety receives reactionary, well intentioned, momentary attention before it’s right back to the race again. ... Make safety part of the race.” Elizabeth Edwards-Appell, Anthropic: “None of us yet know how to make sure these things stay under human control and/or take actions only aligned with the wellbeing of humanity. This is, objectively, an insane and suicidal thing to do, especially without any international governance measures in place.” Patryk Grzelak, Anthropic: “Developing technology shouldn’t require a leap of faith for our continued survival.” Thomas Neil James Shadwell, OpenAI: “We must all seriously consider the lead-time required for defenders to be ready for the frontier.” Leo He, Google, “The core principle is simple: no AI system, however capable, should be the sole judge of whether its successor is safe to deploy. Fund the independent evaluation and oversight that makes that principle real.” Fin Moorhouse, DeepMind: “When scientists tested the first nuclear reactor in 1942, they feared a runaway reaction — so they built control rods to govern its speed. We need control rods for the intelligence explosion.” Andy A. Neals, Anthropic: “Let’s not let our future selves look back and ask why we didn’t take action when we had the chance.” These are all sobering assessments from those closest to the action, and we should heed them. The comment that lingers most in my mind is one by Ali Egal of Anthropic: “Think of all we love.” Just so. Pace the frontier.