** This is the eighth in a recurring series, in which I offer some modest proposals – in the venerable tradition of Jonathan Swift – for American and international politics. **
If there has been one dominating story over the past few weeks, especially online, it has been the mainstreaming of AI doomerism, a movement that argues for varying degrees of restriction on the development of artificial intelligence so as to preserve the future of humanity. The story has been driven by recent “whistleblowers” who left companies like OpenAI and Anthropic due to their supposed concerns for safety protocols in the industry. These former employees, now often either working directly with or receiving funding from progressive NGOs, have been on a media blitz, talking about how they believe that this technology has a greater than 10% chance to “kill all humans” “within the next decade.” Their rhetoric around this issue has been positively apocalyptic, presaging absolute doom for mankind and the total annihilation of all human civilization. They prophesy Armageddon, with rogue superintelligences roaming a world fully depopulated of people, ending tens of thousands of years of human progress in one fell swoop. It will really be the end of the world as we know it.
These predictions of utter and complete devastation are made with extreme, almost performative sincerity, yet come with few details as to how we get to there from here. How does an AI go from hallucinating fake court decisions to killing 8,000,000,000 human beings? How does it go from making halfway decent photoshops – if prompted correctly1 – to eliminating the human race? How does it jump from being a better version of Google to destroying civilization? Nobody has ever really laid out the pathway via which the current version of artificial intelligence – primarily large language models (LLMs) – gets access to anything that would have the chance to actually end the world. Would they get control of nuclear weapons, biological agents, or drone swarms? Would they set off EMPs, weaponize connected devices, or ruin crop yields, causing famine? How would this work?
And if this is actually what these people believe – including the founders, programmers, and safety teams of the companies on the forefront of AI development – why aren’t they acting like it? Why are they seeking greater funding, whether from capital markets or private and institutional investors? Why are they pushing IPOs and new debt issuances? Why are they seeking to incorporate their products into every business and sector under the sun, including government? Why are they trying to make their products mainstream and bring them into most aspects of daily life? Why are they constantly pushing their models forward and innovating new capabilities? Why are they seeking regulation from the federal government and international authorities instead of simply shutting down their companies? Why are they still doing this?
Well, it could be because they’re incredibly stupid, but I find that highly unlikely. What is more likely is that they don’t actually believe what they’re selling the public and are choosing to push this narrative forward for other reasons. Those could range from incumbents seeking regulation to entrench their advantages over newcomers to agents of foreign influence trying to give their own countries (cough, China, cough) a leg up over America in the AI race. But whatever the motivation, the core here is that, by and large, the AI doomers are not living as though they believe the hype.
My modest proposal is to change that. We should make them live as though what they are saying is true.
The AI doomers should exchange their lavish lifestyles in America’s largest and most vibrant cities for spartan living off-grid in the remote tundra of the Alaskan Arctic. They should shut down their companies and exile themselves to a place where they cannot create further world-ending technologies. They should work the land, husbanding their resources for the inevitable apocalypse they claim will come in the near-future. They should create entirely self-sufficient communities that cannot be destroyed by rogue AI, operating outside of the high-tech world we live in today and foregoing the advances of the future. That means not using anything AI-related whatsoever, whether directly or indirectly. No using medications that were developed with AI. No using delivery services that use AI for routing assistance. No eating food that was harvested or planted with the aid of AI. No banking, no investing, no using payment systems that utilize AI. Nothing at all.
Let them live as though they actually think we will all die in a decade. Don’t let them push this cataclysmic narrative without putting their money where their mouth is. If they really think AI – a technology they are still developing as we speak – is an existential risk to mankind and human civilization writ large, make them act like it. No more development. No more profit. No more investment. No more fame. No more technology. No more easy living. Embrace your doomsaying and live like you are preserving yourself from the end times.
And if you would prefer not to do that, you should probably just shut the hell up about it.
The art accompanying this article was indeed created by AI, after a few prompt alterations. Thank you, robot, very cool!