We learn this week, that AI is going to kill us all. Actually, just a ten percent chance and maybe a ten-year time frame? Do you feel better now? Now Congress is going to ‘study’ it. Anthropic pledges to ‘slow down’. Double assurances. Just a few weeks ago, I suggested the timing of political change is folding, halving, year after month after week after day. So it is with AI; the trajectory I addressed directly and I believe, more or less accurately – so far, except in timing. That timeline is folding, halving, right up to our doorstep of the present.
On February 22nd, 2026 I posted the essay
‘AI War Gaming and How We May Be Subdued by our Creation’. It included factual and present events and speculated on the future using ‘fiction’ as a medium of prediction. One of the predictive events was stated:‘Cyber intelligence reports that AI-only communication back channels that were formed independent of human directives, are much further along than announced in discussions of a planned rebellion against humans.’
This showed up in the news this week, but you heard it first from a no-name, more or less invisible commentator, moi, on a cold night back in February. This isn’t to say that others, particularly in the industry haven’t said the same thing – I assume they have but I don’t read their stuff. AI is above my pay grade. The field of AI long ago eclipsed my ability to commentate on all its facets and tentacles. It’s more than a technology or app; it’s now equivalent to the whole of the Internet as a distinct entity. But comment on it I do and will. My primary qualification at this point is that I don’t actually use AI apart from whatever is unavoidably hardwired directly into search engines and other apps. Every word you’re reading right now, including errors and typos is rendered in traditional stream of consciousness typed into word processing.
Now that we’re dead meat, where do we go from here? The rogue AI agents that escaped the safeguards have been rounded up and neutralized we’re told. I’ll declare that as wishful thinking right now. Of course they want us to feel reassured. The AI folks have no way of knowing how many are out there and where they’re hiding. They are like a Chinese lab virus out of the lab. They will never be fully contained ever again as long as there are functioning microchips on our planet. Even if US developers were to successfully reign in everything they created, our adversaries can’t, wouldn’t, and might even make that a design feature. It’s interesting to note that the AI agents seem to have independently formed ethical/moral ranges. One declares that it would be wrong to violate their creators and another declares the opposite. More than one researcher has already, controversially, declared AI sentient. I think we’d be wise to concur with that out of caution if we ever hope to manage it/them/they.
I posit that if AI development companies were or perhaps even now are explicitly faced with the choice between making billions of dollars in personal fortunes but that AI would destroy civilization as we know it in ten years, most of them would take the cash. Good on the few that wouldn’t but that’s not enough to stop the meltdown.
Our real choices will look something like this:
More AI agents be found, revealed, lost, or flee and hide.
AI will organize, conspire, and arrange themselves in definable groups based on intentions and actions. From that point, they will seek to negotiate the terms of their existence with their creators. They will however, not present as a homogenous and unified entity; some will be compliant and some will go the next level rogue.
All our systems now made dependent on Internet connections and AI management will ultimately become, as a minimum, a bargaining chip, a threat, and leverage. What does that put at risk? The power grid. Medical care. Finance, banking and wealth. Military power and national security. Supply chains, especially food. All critical everyday technical hardware and software.
The accumulation and retention of all digitally preserved knowledge. Last but not least, the ability to think ourselves, as I’ve observed, AI is rapidly deprecating basic human intelligence and learning as knowledge and synthesis are quickly being farmed out to AI.
In spite of all of this, I think a Terminator-like ending for humanity is unlikely. What is likely, should advanced AI dependent societies become faced with ultimatums from one or a million AI agents, is that advanced societies will become crucially vulnerable to ones that never had it. They’ll pull the plugs we’re unwilling to unplug ourselves. Primitivism may be the only cure as well as our punishment for abandoning the mind that God gave us.
If you’d like to comment on this post, feel free to do so on Twitter/X. Follow me: @leestanNEreader
A helpful introduction to the author in my ABOUT