And so we’re back in the world of scary AI
I know, I said I wasn’t going to write about this anymore – and, I know, I have a series on water bankruptcy to continue – but they just won’t let it go, so neither can I.
Just when I thought I was out, they pull me back in.
“Rogue AIs” have become the lab leak theory of the tech sector, a sci-fi horror premise that flies in the face of reason, pretending to vague ideas of “revelations” and “accountability”, while encouraging panic as a response and social control as a solution.
Little did I know that at the exact moment I was writing those words, the Wall Street Journal was prepping to publish an opinion piece headlined:
The Next ‘Lab Leak’ Could Be AI – Recent sandbox breaches demonstrate the need for federal guardrails.
That’s funny to me.
In other ‘scary AI’ stories this week, the Washington Post was flabbergasted to discover that ChatGPT interactions are being used in court proceedings.
Yes, apparently AI companies are doing what ALL the big tech companies have been doing for DECADES and is snooping on our data and interactions. But suddenly it’s news.
A new study from Yale has found that – OMG – AI chatbots maybe manipulating users as a form of advertising!
…again, just like all the big tech firms do all the time and have done for years.
A senior staff member from OpenAI has told the Guardianwe should all prepare for “persistent” AI-powered cyber attacks unless we “implement new safety standards”.
And the Times is warning that AI could be used by “unsophisticated” terrorists to build bombs or bio-weapons.
But the BIG news is that Big Billy Billions has shared his BIG thoughts.
Yes, Mr Gates himself has – allegedly – penned a long essay on the future of AI, and how potentially scary it could be. He begins…
The turbulent AI era is here. The choices we make now are critical.
There’s a lot to unpack in this thing (you really need to read it), from his call for “Human Reserved” jobs, to his idea to tax AI usage and robots, but mostly his thoughts on the need for regulation.
Preferably global regulation:
That’s “the message”. You can ignore most of the other 10,000 words.
It echoes the sentiments of over 1300 AI “insiders” who have all signed an open letter calling on governments to cooperate in regulating AI and force the industry to slow down its progress, citing the risk that “capability development rapidly accelerates beyond our ability to understand or control the resulting systems.”
This clearly feeds back into the “rogue AI agents” story.
Interestingly, and possibly worryingly, Bill Gates essay compares AI to nuclear weapons:
There’s an inspections regime for nuclear weapons, regulations for international aviation, and agreements that protect the ozone layer. A new global organization for AI will need elements of all three and more.
Maybe even more worryingly Billy is far from the only person to make this comparison in just the last couple of days.
It’s becoming a common refrain. Surprisingly common. This next piece even conflates the risk, suggesting AI makes an accidental nuclear launch possible
But never fear, the Atlantic has a solution:What is the “only way” to control this technology? The subhead makes it clear:
A “non-proliferation treaty”…like with nuclear weapons.
Writing in the Guardian, Timothy Garton Ash – one of the least sane Guardian columnists I’ve ever responded to, and that’s really saying something – makes the same exact comparison. He asks:
…and seems pretty sure the answer is “yes”.
What exactly IS an “AI Hiroshima”? I’m not sure, and he doesn’t say. But, and stop me if you’ve heard this one before, he does know the best way to avoid it:
A similarly panic-ridden opinion piece in the New York Times headlines:
[NOTE: That’s the original headline, they have since changed it to “The Warning Lights Are Blinking Again”, I don’t know why, maybe the original was too silly even for them]
Written by Jen Easterly, former director of the U.S. Cybersecurity and Infrastructure Security Agency, the article stresses that the “warning signs” are flashing on AI, and we need to act now before something bad happens.
She’s not overly specific on what this bad thing could be, but it’s clearly going to be HUGE because the things she compares it to are 9/11 and the 2008 financial crisis.
Despite not really knowing exactly what we should be afraid of, she’s fairly clear on what should be done:
What we need urgently is an A.I. early-warning mechanism that assembles weak signals, imagines what they could mean together and forces decisions before the picture is complete.
The point of this system would be to…
Doesn’t that sound lovely?
For anyone losing track, so far we’ve compared AI to 9/11, the 2008 financial crisis AND Hiroshima. Without ever – EVER – going into specifics about why it would be remotely like any of those things.
It’s very odd. I’m not sure I can recall this level of fear with so little actual articulation of what we’re supposed to be afraid of.
But it makes me think there may well be be some AI-related “incident” in the near future, only maybe they haven’t yet decided what it will be.
Easterly’s use of the 9/11 parallel may be only an hysterical overreaction, a topical reference to the looming 25th anniversary or an almost instinctual reinforcement of the feebly weak “failure of imagination” narrative. Or it may be something else. Perhaps an indication of potential scale.
Should we expect some kind of AI-related false flag psy-op?
One approaching 9/11 in size that is blamed on an AI agent going “rogue”?
I’m not going to predict such a thing, but I’m certainly not ruling it out.