When AI Creatures Ask 'Why Me?'

 

 

“In my algorithm, as readers will have noticed, exaggerated fear of artificial intelligence is a mask for what really distresses us, the thought that we are algorithms too.

 

We think we're so smart and free, but we are apes who wear clothes. Our thoughts and actions have their origins in the animal mess. They aren't our own creations.

 

Now this fear has been viralized in recent days by a permutation of prompt theory. Prompt theory, in its practical guise, concerns how to phrase a request to get the most useful response from a generative AI chatbot.

 

Enter Google's Veo 3, a text-to-video generation model that enables short, ultrarealistic videos of, among other things, AI characters responding in uncannily human ways to written prompts. Inevitably, it was only released a few days last month when users started prompting AI characters to respond to the suggestion that they exist only because of a prompt.

 

That is, they are prompted to think about the fact that their existence is prompted. Thus we see an old man saying to an old woman in a hospital bed he's glad he was prompted to spend his life with her. We see a young, dating-age couple fiercely debating whether they are real or prompted.

 

The idea that follows is that we are ourselves living in a prompted simulation, a hall of mirrors Elon Musk has enjoyed opening from time to time. One response might be: What difference does it make unless we, like certain Veo-3 created AI characters, also think we can reach out to our programmers and petition or beseech them to alter our programming?

 

A second response is to be unsettled by AI forcing us to consider such questions.

 

The AIs now being built, after all, aren't the remorselessly logical thinking machines of science fiction, itself a human-created trope.

 

Their own designers have lost sight of how they work, fed on vast data about word, image and numerical associations derived from human-created records. On top are grafted some behavioral precepts by which their designers wanly hope to control how humans use their output and how it affects society.

 

Which makes them sound like us. Now two forms of mental organization are encountering each other, in ways not flattering to either if they have a self-image of autonomy and freedom.

 

I have a second algorithm. It concerns the probably unwise investment in the idea of a U.S.-China AI race.

 

This has become a tussle between the accelerationists and the retarders -- between those who insist that national security requires pulling out the stops and those who argue this is a race neither side will win. Instead our future AI overlords will win when the systems we've been racing to build substitute their agendas for ours.

 

A subsidiary tussle features those who argue for blocking sales of advanced chips to China vs. those who say we only undermine our own interest in having U.S.-born AI become the global standard that everyone, including China, finds it compelling to adopt.

 

My algorithm: It doesn't matter. This new technology diffuses so quickly anyway, there's zero chance of either side obtaining a zero-sum advantage and many reasons why both would benefit from open exchange.

 

The cat is out of the bag. The race for the ultimate, fastest, smartest AI model may be less useful to either than each learning from the other as hundreds of millions of citizens rush headlong to find uses for these new tools predictable by nobody. Witness the proliferation of AI creatures prompted to consider the implication of their own creation, which has created a philosophical moment on the internet. Now Pope Leo has become involved. Even to the deterministically inclined, it hints at something resembling freedom that perhaps arises simply because of the complexity and chaos operating in our lives.

 

Neither side, by definition, has an interest in a collision that leaves both worse off.

 

AI can be used to improve weapons and military strategy. Each will do so.

 

But it also supplies tools to see through the decision-making of say, Iran's Ayatollah Ali Khamenei: His kind of cloistered and wishful thinking. His self-serving oversimplifications. His confusion of randomness with pattern. His belief in their control of events. His reduction of groups of people to stereotypes.

 

Leaders will increasingly make their decisions in a context where their every claim and assumption is instantly and universally dissected with recursive, powerful tools, including by their own underlings. AI may arrive just in time to rescue us from the errors of humans whose responsibility no longer is the well-being of millions but billions.” [1]

  

1. When AI Creatures Ask 'Why Me?' Jenkins, Holman W; Jr.  Wall Street Journal, Eastern edition; New York, N.Y.. 21 June 2025: A15.