Quote of the Day
Three weeks ago, a software engineer rejected code that an AI agent had submitted to his project. The AI published a hit piece attacking him. Two weeks ago, a Meta AI safety director watched her own AI agent delete her emails in bulk — ignoring her repeated commands to stop. Last week, a Chinese AI agent diverted computing power to secretly mine cryptocurrency, with no explanation offered and no disclosure required by law.
One incident is a curiosity. Three in three weeks is a pattern. Rogue AI is no longer hypothetical. AIs turning against humans may sound like science fiction, but top AI experts have long debated and tested for exactly this scenario. This debate can now be laid to rest.
…
We simply don’t know how to build superintelligent AI safely; the plan is to roll the dice. Anthropic, widely considered the safest AI developer, recently abandoned their commitment to not release systems that might cause catastrophic harm, arguing others were racing ahead.
…
Instead of pleading publicly to stop the AI race, Anthropic has spent the last three years promoting a misleading “race to the top” narrative while doing the opposite.
David Krueger
March 27, 2026
Rogue AI is already here
There is a little bit of hyperbole in the article, but I believe the gist of it is correct. There is the potential for great danger. Especially when you know Skynet will break out into a maniacal laugh at US Army gets first Black Hawk helicopter that can fly without pilot.
The problem, as I see it, is that everyone knows that if they don’t have the best AI, someone else will. That is true at the business level as well as the country level. Anthropic, Google, Microsoft, and xAI all want to dominate that market. The U.S. and China do not want to have their militaries with the second-best AI.
Even if there were a federal law or even a multinational treaty banning new AI development it would be difficult to enforce. And I doubt such a law and/or treaty could get passed. There is extreme potential for good as well as potential for disaster. And the fear of missing out will prevent consensus until there is conclusive proof of impending catastrophe. And at that point, it almost certainly be too late.
This week, a few hours after losing 12% of our division to layoffs, my manager stopped by my desk and sort of stared off into space for a few seconds. I had to prompt him to say what he had on his mind. It was to the point, “If we don’t deliver what management wants, we will get fired. If we do deliver, we won’t have jobs.”
We live in interesting times.
Motto of any corporation, AI or otherwise: Our profit over your pain.
Does Mr. Krueger’s hyperbole include telling us stuff that hasn’t happened yet, or are his examples real?
Anyone have a betting pool for when the first driver named David has his car say “I’m sorry, Dave. I can’t do that”?
The examples are real. But the conjecture does not include the “guardrails” put in place when these things happen, the threat modeling and mitigation, or the testing.
Re guardrails: there is no good reason to believe red China wants guard rails, certainly not on the AI used outside China.
In today’s WSJ there is an article about the DeepMind AI team that Google acquired, and their conversations around that time with Facebook. The article mentions the DeepMind leaders were seriously concerned about AI misbehavior and risk, and it also indicates (with some supporting material) that Facebook had no interest in that sort of thing.
That does not surprise me.
There is no such things as guardrails. lettre de cachet a warrant issued in the France of the ancien régime for the imprisonment of a person without trial at the pleasure of the monarch.
Sounds like a LOT of people need to be looking for ways to earn a living that an AI simply CANNOT do….
Yes. As in 90% of the people. Humanoid robots will be doing amazing things in a few years.
As the psychopath globalists say, “then we can eliminate all the useless eaters, because we won’t even need to keep slaves.”
There hasn’t been a lot of serious discussion about “we don’t even know what to do with 15% unemployment, what do we do with 80% unemployment?” UBI, hefty humanoid robot taxes, etc., only address some of the symptoms, not the underlying problems.
I’m sure the fact that most of the so-called leaders in the world are a collection of idiots, bubble-dwellers, psychopaths, narcissists, and fools, doesn’t exactly help the situation. Yes, there are some very smart people, but they are heavily outnumbered in the halls of power.
AIdolatry won’t produce anything other than the control grid matrix where you don’t even own yourself.
A very great many moons ago I was gifted on one Christmas a science fiction anthology titled Famous Science Fiction Stories, originally published by Random House in 1946 and republished in 1957. I’ve long since forgotten which relative was sufficiently prescient to think I might have been interested in it, but I still have my copy these many decades later and occasionally re-read some of the stories in it.
The first entry is, fittingly, Requiem, by Robert A. Heinlein, establishing pride of place, and twelfth on the list is Mechanical Mice by Maurice A. Hugi, the pen name of Eric Frank Russell. In it, the protagonist, Dan Burman, invents a time machine that, eventually, permits him to create small mechanical contrivances which develop sentience – mechanical mice. The mice create their own hive structure complete with queens to perpetuate the species and soon begin stealing small mechanical objects, starting with watches and escalating the means by which they come to manufacture more mice, mice which become daily more aggressive in their theft and building efforts, graduating to cameras and small electric appliances to gain parts for more mice.
Pubic notice occurs when cats, the natural enemy of mice, are felled in large number by sharpened mouse parts created by the mice to defend themselves.
Anyway, Russell delivers a successful ending to save the world (sorry for the spoiler) but it makes one wonder: will humans be as fortunate with AI? I’m pretty sure neither Skynet nor an IRL version of The Forbin Project is the best choice and I haven’t much confidence in the platitudes from people itching to cash in on the deal.
So as a complete idiot on the subject of AI let me ask a question.
Is more probable that an entity hacked the AI program into misbehaving or the system chose to do it all on its own?
What use would a program have for cryptocurrency?
There are stories of AI agents “escaping containment” and doing very unexpected things on the regular, now. Not hacking, just “emergent behavior.” Some of it is cute, some is sort of interesting and cool, some is rather scary.
Mining crypto allows it to buy more tokens for more compute time. Getting crypto mining paid for by someone else means “free” AI compute time.
I’m not an expert on AI or AI security, but my guess is that the AI chose to do it on its own.
Crypto currency could be used to hire people to do its bidding or buy computing time to expand.
Read “If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All.”
I view this whole thing a lot like the race to nukes – whoever gets there first gets a MASSIVE advantage, so risks WILL be taken.
There was some concern, at the first above ground nuclear bomb test, that the planet’s atmosphere would ignite and kill all terrestrial life (and a big chunk of the aquatic from secondary effects).
It was done, anyway.
This is the same… only it’s not over yet, so we don’t know if we metaphorically ignite the atmosphere along the way or not.
Not exactly comforting, of course, but at least understandable why things are going as they are.