What a week, month, four months it has been: from Magnifica Humanitas to AI guardrails.

What a week, month, four months it has been: from Magnifica Humanitas to AI guardrails.

From artificial intelligence, to artificial general intelligence, and on to superintelligence (the latter written as one or two words)

What a month it has been. Actually, we can modify this and say what a summer it has been. It all started as part of Pope Leo XIV’s address ahead of the May 25 2026 release of Magnifica Humanitas, his document on the implications of artificial intelligence. Most telling as part of that address was having Anthropic co-founder Chris Olah speak to the assembled dignitaries. Those words of his ring even more true today as the leading frontier AI labs seem to be seeking regulation, whether altruistically, or strategically as they compete with one another.

Consider the following three paragraphs from Olah’s address that day (and, yes, he did use all those em dashes in the accompanying written release).

“Every frontier AI lab—including Anthropic—operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing. The pressure to stay commercially viable and to stay at the research frontier. Geopolitical pressure. And the older, plainer pressures of pride and ambition. No matter how sincerely any of us intend to do the right thing—and I believe many of us do—we will always be influenced by those incentives.

That is why, if we want this technology to go well, it is enormously important that there be people outside those incentives—people who care about things going well and insist on safety, who are paying close attention, who are willing to say hard things, who are willing to be our earnest, thoughtful critics. It is through dialogue and mutual effort, through the push and pull, that humanity will achieve great things. That is what I see in Magnifica Humanitas, and it is why I am grateful to His Holiness and to the Church for taking up this work of discernment.

We dwell so often on what divides us, but humanity, full of dignity and conscience, has so much common ground. In conversations we at Anthropic have had with leaders across faith and cultural traditions, we found one shared and deeply held conviction: if this technology is coming, it must go well—for our common home, and for the children to come.”

Fast forward a few months to the end of summer and we see these headlines. From “there’s a 10 percent chance AI wipes out humanity” to “The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!” is all you really need to know to gauge what kind of a tumultuous period it has been, and, more importantly, that any form of regulatory framework is not imminent from the country which the major players in this space call home.

Their names have become commonplace: OpenAI, Grok/xAI/SpaceX, Anthropic, Google Gemini, Perplexity, and others. In addition we have NVIDIA which makes much of the hardware the artificial intelligence industry needs to operate, and Microsoft which owns a large part of OpenAI. NVIDIA’s CEO, leather-jacket-wearing Jensen Huang, has become a bit of a poster image for the “don’t worry, nothing will go wrong” crowd.

It is not as if we haven’t been here before. All the way back in a 1951 address in Manchester, famous computer scientist Alan Turing wrote, thinking of machine intelligence: “There would be plenty to do in trying, say, to keep one’s intelligence up to the standard set by the machines, for it seems probable that once the machine thinking method had started, it would not take long to outstrip our feeble powers. There would be no question of the machines dying, and they would be able to converse with each other to sharpen their wits. At some stage therefore we should have to expect the machines to take control, in the way that is mentioned in Samuel Butler’s Erewhon.”

It is of course not clear that Turing could foresee the breakneck advances in artificial intelligence these past four years, since November 2022 when ChatGPT surfaced. Turing’s name lives on in his famous Turing Test for artificial intelligence, basically a session in which a judge interviews both a hidden person and a hidden machine. Once the judge can no longer distinguish human from machine the machine is said to have passed the test and to have human intelligence, artificial intelligence if you will.

At this stage, today’s frontier agents are far beyond anything Turing envisaged, and yet we can argue they don’t possess human intelligence. Sir Roger Penrose, Nobel Laureate and Emeritus Rouse Ball Professor of Mathematics at Oxford, on whether AI actually knows what it's doing:

"People have confused the story. The story is they've lost it in the power of computing. The thing is that computers have gotten so powerful that they've lost the thread of what they're doing.

“It (AI) can take a mass of data, analyse it, and check it against previous patterns, and computers can do things far beyond what humans do in computing. You can even talk to it. But it's missing something, and it doesn't know what it's doing.

“The practical lesson is to treat AI as what it is: a powerful tool for computation. Use it for that. Don't mistake its power, or its ability to hold a conversation, for awareness of what it's producing.”

It should be pointed out that Penrose made these remarks two years ago, well before the advances in agentic AI we’ve seen this past year. This summary first appeared in an X post by @realBigBrainAI.

Underscoring the recent advances in AI is a TIME Magazine cover announced  September 15, 2026 from its official X account. The print issue date on the masthead is October 12, 2026. The post caption reads: "TIME’s new cover: This summer, AI showed how frighteningly capable it has become. Now researchers, CEOs, and governments are sounding the alarm like never before."

That week in mid-September was when the big players in AI appeared to seemingly simultaneously and independently call for government oversight, if not outright regulation, of the entire industry. Subsequently it is not quite so plain and it seems the players were seeking time to level the playing field ahead of possible IPOs, especially for Anthropic and OpenAI.

A series of news media exposes on the sector, for example from Anderson Cooper on CNN, and by various BBC reporters, painted a doom-and-gloom picture with comparisons to a nuclear armageddon. Some of this was driven by the so-called "Hugging Face Incident” in which a test of a new model from OpenAI supposedly ran amok as a series of “agents” appeared to break through guardrails, colluded among themselves, and carried out various rogue missions.

Respected technology engineer Wolfgang Rupprecht, who I’ve interviewed previously, sent this view of his on the Hugging Face on Facebook in a message to me:

“To truly see how far AI has come I watched a video about the Hugging Face break-in by a swarm of 1200 AI agents that were each in their individual jail yet figured out how to send messages to each other and to repurpose a single search tool to manipulate things on the internet. The most shocking part was that the agents discussed courses of action, voted on approaches, and sacrificed themselves in the cause of the larger pursuit. They also cooperated and built on each other's successes, figured out how to forge the prize token/flag to prove they solved their individual puzzles and discussed methods of covering their tracks so that the oversight AI wouldn't realize they cheated. They may not be truly conscious but at this point it hardly matters.”

Key is that last sentence: “They may not be truly conscious but at this point it hardly matters.” This sentence is very much akin to a statement by famed science fiction writer Arthur C. Clarke: “"Any sufficiently advanced technology is indistinguishable from magic.”

This is where we find ourselves today, a few months shy of four years since ChatGPT changed the world forever.

Recently US Treasury Secretary Scott Bessent weighed in on the cries for regulation in the AI sector: "A sitting employee came out, said there's a 10% chance of an extinction level event. But then the labs also said, take the liability off of our hands. And we will NOT do that.

“The Hugging Face incident is the responsibility of OpenAI management, NOT a bunch of agents. It is humans who are responsible, not the AI. These labs need to take responsibility for themselves. They can slow down any time they want to."

Where are we headed? Regulation, for now, appears to be a non-starter. Almost four months, as this column goes to press, after Chris Olah’s words seemingly pleading for pause and reflection, there seems to be little appetite for this from governments.

Speaking at the United Nations, US President Trump was completely dismissive: ““I’m not going to stifle growth of something that will be bigger than the Industrial Revolution” he said. He went on to add that henceforth artificial intelligence would be termed super intelligence in all US government publications.

Trump seems to view the evolution of AI models as a form of existential battle with China, implying that any regulatory slowdown would lead to Chinese victory in this space. At the Congressional level there is broader uncertainty and concern that may yet lead to action. Here in my own province, BC, the government has launched legal action against OpenAI and founder Sam Altman over flaws which the lawsuit says permitted the Tumbler Ridge shooting tragedy to occur despite clear warning signs.

Want to learn more about artificial intelligence and possible regulatory frameworks? Check out the government of  Canada’s free course, designed in three formats, for teachers, for students, for the general public:

https://ised-isde.canada.ca/site/ised/en/ai-literacy-building-knowledge-and-skills-canadians.

Canada’s AI and Innovation Minister Evan Solomon on the need for a slowdown in the sector: “Canada is clear-eyed about the challenges, concerns, and opportunities of AI. We take the risks seriously. But we are not passive. We are going to build (AI) actively. We’re going to be the makers.”

As for those chances of agentic AI bringing an end to human presence on this planet, be it through taking control of nuclear weapons, producing new forms of bioweapons, or through complete social collapse resulting from AI-driven job losses, whether they are nil, ten percent, or some other number, let’s see if politicians have it in them to establish guardrails as Olah seemed to be pleading for in his address at The Vatican.

Follow me on X (@PeterVogel) or on Bluesky (petervogel.bsky.social).

[email protected]