There is a limit to the amount of existential dread a human mind can comfortably hold before the machinery simply jams. I’m hoping a gun to Mr Cheeseface’s furry little head can help unjam the works and shake us from our ontological shell shock.
Tell a friend that a leak from a lab in California might lead to an artificial intelligence architecture capable of destabilising global financial systems, and you’re more likely to be asked for help with a Wordle than recommendations for a good bunker.
Or tell your mum that an uncontained recursive intelligence model could extinguish all organic life within a generation, as a fresh batch of whistleblowers spent this week trying to explain. Go on. See how that goes down. Look, I love your mum, we all do, but let’s at least acknowledge that her Facebook updates are all Google searches in the wrong box.
And it’s not just our mums. We’re all famously hopeless at scale.
“The death of one man is a tragedy; the death of millions is a statistic.”
Who said that, then? No points if you guessed Joseph Stalin. The earliest documented version appears in Kurt Tucholsky’s 1925 essay Französischer Witz, spoken by a fictional French diplomat; the Stalin attribution surfaced later without evidence. Yes, technically Stalin may have said it, if you’re that keen for points, fine, but it doesn’t reflect well on you; people have picked up on this kind of shit and are afraid to tell you. It’s only right you should know.
Anyway, regardless of who said it, it’s an acute piece of cognitive psychology. When a threat expands past a certain threshold, it ceases to feel like a threat at all. It disperses into the background and feels like the hum of a crap fridge or tightness in the chest on a Sunday evening.
Which brings us to a scruffy, black-and-white mongrel named Mr Cheeseface.
In January 1973, National Lampoon published what would become one of the most famous magazine covers in publishing history. I didn’t see it at the time; I was three and British. But I can read Wikipedia, so I now know everything about everything.
Photographed by Ronald G. Harris under the art direction of Michael Gross, the iconic Lampoon cover showed Mr Cheeseface staring nervously out at the reader while a blued-steel revolver was pressed flush against his temple.
The cover line: “If You Don’t Buy This Magazine, We’ll Kill This Dog.”
To magazine geeks like me, that’s like a da Vinci sketch pinned to an actual da Vinci. And it worked really well, not just because it was outrageous, but because it possessed a shrewd understanding of human emotional mechanics.
In the early 1970s, newsagents’ shelves will have been overflowing with glossy covers documenting distant horrors: the slow crawl of the Vietnam War (before it got its Hollywood makeover in the 80s), nuclear proliferation (at a time when the superpowers were sane enough to actually be working on legislation to limit proliferation), and industrial collapse (hey, it’s no coincidence that 1979 gave us the first Mad Max movie).
Readers had developed a thick hide of moral fatigue.
The Lampoon team realised that if you want to pierce that hide, grand moral imperatives won’t cut it. You’ve got to collapse the entire catastrophe down to a single, absurdly manageable quandary.
If we are to make any sense of the warnings echoing out of Silicon Valley this week, we need to keep things simple. The modern AI safety debate has spent years drowning in its own vocabulary. Alignment. Autonomous systems. Deceptive behaviour. Recursive self-improvement. Instrumental goals. Existential risk.
These are useful terms if you are sitting in a room with people who already understand them. Or those people who are paid to understand them, don’t, but can’t face the idea of losing the salary so do as much remote work as possible and move the mouse with their feet while laid on the sofa so they look busy on Teams.
To everyone else, the AI jargon can make the whole thing sound like a plot outline for a mid-budget science-fiction film. Something you might watch on a Friday night after the pub, if pubs still exist when this is published.
If the scenario warned of by this week’s whistleblowers unfolds – if an unchecked race for market dominance produces a system we cannot recall or contain – Mr Cheeseface dies. For real this time. Again.
And “We” – the ones holding the revolver – is not some shadowy cabal of rogue software developers. It is all of us, sitting quietly on the train, scrolling through AI slop and alternative realities presented by swarms of state-sponsored bots, while the safety guardrails are unpicked behind closed doors.
Crucially, though, it isn’t just the dog.
Your mum? Dead.
Brian Cox? Dead.
Sir David Attenborough? Dead.
Come on! Attenborough! Surely Attenborough is a red line?!
And don’t bother doing the mental maths on the rest of the world – the billions of strangers whose names you’ll never know, those you might quietly suspect you care about slightly less than a pigeon with a gammy wing you once found and took home.
They are dead too. Including that pigeon.
Hey, it’s not just dogs, Cox and our mums. Your cat? Dead. Unless, of course, it manages to broker a separate peace with the autonomous agent swarms and sell out the household in exchange for fresh salmon. Cats will do that.
At this point, a cynic enters the chat.
“The whole AI threat is a global deception built by technology billionaires as some kind of Project Blueish Beam to distract while they all pass around a trillion-dollar debt parcel during a frenzied search for the stock market’s final form as a financial Godzilla Pokémon.”
Well, isn’t that just changing the livery on Air Dystopia? We’re still trapped in Economy on a long-haul flight to hell, someone in 14B has brought a laryngitic baby seal support pup, and the in-flight entertainment is stuck in a 13-second repeat fever dream of Mr Bean locked in a grand mal seizure.
Whether the cockpit is being operated by a rogue hyper-intelligence or a roomful of gilded sociopaths engineering the ultimate speculative bubble while the fuselage falls apart, the destination’s the same, right?
The crash landing doesn’t get any comfier just because the pilots are doing it for shareholder value rather than autonomous malice.
The joke on the cover of National Lampoon was, of course, a bluff. Although, according to Wikipedia, Mr Cheeseface, in a piece of cruel historical irony, met a tragic end three years later on a farm in Vermont at the hands of an unidentified hunter.
Regardless, the psychological principle behind it remains the same. We cannot care about eight billion people at once. We might ache to really, fully, truly connect with them all, but we lack the emotional bandwidth.
We can, however, care about what fits in a frame we can work with.
We do not need to master the intricacies of neural network architecture, or untangle the grifts of Silicon Valley boardrooms, to understand what is happening in the news this week. We only need to recognise our own capacity for wilful blindness when a risk gets too big to look at directly.
Choose regulation. Demand the guardrails. Save the dog.
- George Hopkin thinks magazines can save the world. Read his here. His magazine, that is, not his world. Heh, can you imagine?