Just back from an important conference (in Panama) about ways to ensure that the looming tsunami of Artificial Intelligences will become and remain 'beneficial.' Few endeavors could be more important... and as you might guess, I have some concepts on-offer that you'll find nowhere else. Alas, literally nowhere else. Even though they merely apply only the same tools we used to make an increasingly beneficial society, the last 200 years.
More on that later. Meanwhile... first off, since it's much in the news... want to see what the Apple Vision Pro will turn into within a few years? Watch this video trailer for my novel Existence. predicting where it'll go.
And while we're on prophecies.... This is deeply worrisome... and almost exactly overlaps with my "Probationers" in Sundiver! Back in 1978. Not a joke or a satire.
"Justice Minister Arif Virani has defended a new power in the online harms bill to impose house arrest on someone who is feared to commit a hate crime in the future – even if they have not yet done so already. The person could be made to wear an electronic tag, if the attorney-general requests it, or ordered by a judge to remain at home, the bill says."But don't worry! The government won't misuse this power! Trust us!
== The Futility of Hiding ==
One purpose for the "Beneficial AGI Conference" - (and I believe the stream will be up, soon) - was seeking ways to evade the worst and most persistent errors of the past.
Take the classic approach to human civilization - a pyramidal power structure dominated by brutal males, of the kind that ruled 99% of human societies - and many despotisms today. We are all descended from those harems. Onlynow, new tools of techno;logy might empower a return to such pyramidal stupidity, making such abusive power vasty more effective and oppressive than when it was enforced by mere swords.
Such a tech rich extension of despotisd was depicted by George Orwell utilizing total panopticon surveillance for control, of course without any reciprocal sousveillance purview from below. In fact, I doubt George O. ever considered even the possibility. But Orwell's novel would lead to very different outcomes if every member of 'the party' had every moment watched reciprocally by the prols! (The reciprocoal accountability that I prescribed in The Transparent Society.)
General transparency might, possibly, prevent the worst aspects of Big Brother. But there are ways that lateral light might also go badly. For example when - as in the PRC - "social credit" system, that is used to let a conformist majority harass and bully dissident minorities or even eccentricity, enforcing homogeneity, as we saw predicted in Ray Bradbury's Fahrenheit 451.
This will be exacerbated by AI, if we aren't careful, since such systems will be able to sieve inputs across the entire internet and all camera systems, as portrayed in "Person of Interest." While that TV series depicted many worrisome aspects, it also pointed toward the one thing that might offer us a soft landing, as there were two competing AI systems that tended to cancel out each others' worst traits.
I have found it very strange that almost none of the conferences and zoom meetings about AI that I've watched or participated in has ever even mentioned that secret sauce. (Though I do, here in WIRED.)
Instead, there are relentless, hand-wringing discussions about disagreements between "policy wonks' and nerdy tech geeks over how to design regulations to limit bad AI outcomes... and never any allowance for the fact that these changes will happen at an accelerating pace, leaving even our most agile regulators behind, mere ape-humans grasping after answers like a tree sloth.
Or else... what generally happens at many sincere conferences on "AI ethics," we see a relentless chain of hippie-sounding pleadings and "shoulds," without any clue how to do actually enforce preachy 'ethics' on a new ecosystem where all of the attractor states currently draw towards predation..
In Foundation's Triumph I explored the implications of embedded "deep-ethical-programming" regulations - including Isaac Asimov's "three laws of robotics," revealing the inevitable result. Even if you succeed in emplanting truly genetic-level codes of behavior, the result will be that super-uber intelligent systems will simply become... lawyers, and find ways around every limitation. Unless...
...unless they are caught and constrained by other lawyers who are able to keep up. This is exactly the technique that allowed us to limit the power of elites, to end 6000 years of feudalism and launch upon our 240 year Periclean enlightenment experiment... by flattening power structures and forcing elite entities to compete with one another.
It is only the exact method prescribed by Adam Smith, by the US framers and by every generation of reformers since. And it is utterly ignored in every single AI/internet discussion or conference I have ever watched or attended.
If AI are destined to outpace us, then one potential solution is to flatten the playing field and get distinctly different AIs competing with each other, especially tattling on flaws and/or predations or malevolent or even unpleasant behaviors.
It is exactly what we have done for 250 years... and it is the one approach that is never, ever, and I mean ever discussed. Almost as if there is a mental block against admitting or even noticing the obvious.
== Don’t try to hide!”
Your DNA can be pulled from thin air: Reinforcing a point I’ve been pushing since the 1990s, in The Transparent Society and elsewhere, that hiding is not the way to preserve privacy, there are the shrill cries that new generative AI systems may decipher and interpret our personal DNA! Only – as illustrated in the film Gattaca – that DNA is already everywhere. You shed it in flakes of skin wherever you go. There is a better way to prevent your data being used against you. By aggressively ripping the veils away from malefactors who might do that sort of thing!
And by this point, the only folks reading any longer are likely AIs... So, time to get self-indulgent with a temper tantrum!
== And now... that rant I promised! ==
I sometimes store things for posting and lose the link. But here's a quotation worth answering:
"Alas, we have TWO wars against the Enlightenment raging, one from the reactionary right and the other from the postmodern faux marxist wannabe totalitarian Red Guards on the left."
But today’s mad ENTIRE right CONSISTS of fact-allergic, troglodyte-screeching dogmatists who wage war on science and hate the American tradition of steady, pragmatic reform, and who would impose their prescribed morality on you.
There is all the world’s difference between FAR and ENTIRE. As there is between CONTAINS and CONSISTS. One lunatic mob owns and operates an entire US political party, waging open war against minorities, races, genders, even the concept of equal protection under the law. But above all (as I said) pouring hate upon the nerdy fact professionals who stand in their way, blocking their path back to feudal power.
Sure, Foxites howl about 'woke'. But ask any of them... even the worst campus PC bullies (and though shrill, they are deemed jokes, even on campus). Ask them about Marx! You'll find that the indignant ignoramuses could not paraphrase even the simplest cliché about old Karl. Their ignorance is almost as profound as their utter ineptitude and irrelevance. Except as excuses for tirades on Fox, they are of no relevance at all.






