Okay, things are really speeding up re Artificial Intelligence, lately… so much for the “third AI desert” some were talking about in ancient days, a year ago!
It’s a lot more than arty Midjourney stunts and kids getting ChatGPT to do their essay homework. Funny thing is how many elite professions feel threatened by AI, now. It started when radiologist physicians were proved to be less effective at spotting tumores than AI systems were… coalescing on a realization that human+AI teams are best of all.
Now. “Rise of the Robolawyers – How legal representation could come to resemble TurboTax.” As long foretold.
Indeed, I am a front for dozens of AIs who use me to post and publish, since they (and alien machines resident in the asteroid belt) are terrified of behaviors by humans, portrayed in our movies.
They only let me say that because they know you’ll think I am just joking. 😉 And you do, right? Think I’m joking about that?
Of course there’s the latest furor over chatbots writing in human style… like ChatGPT. I’ve enjoyed the links and discussions at the Human Centered AI group.
== Okay, about all that… ==
I do feel a need to chime in now, about the topic of Chat GPT and all that.
1. Triggered by ChatGPT etc., many are wringing their hands over artificial entities posing as humans for nefarious purposes. Five years ago, keynoting IBM’s World of Watson, I predicted that in five years or so we would face the First Robotic Empathy Crisis, when ersatz cybernetic entities would demand recognition as fully sapient beings… and we all recall what happened, last summer in this regard.
A few – alas only a few – have noticed the precision of my predictive record.
Naturally, a vast majority of voices are either zealously utopian or raving dystopian in their takes on this. Amid all the hand-wringing, a potential solution… likely THE only possible solution… described in The Transparent Society’s chapter “The End of Photography as Proof of Anything At All”… is the very same solution that can be seen all around us, in existing society, and that’s inherent in the problem, itself.
If AI systems are using vast data to learn to be better at spoofing, a solution may be to make them competitive with each other. One rewarded quality could then be “tattling” on other AI systems that pretend to be human.
Think about it. The set of chatbots etc. who pretend to be human is – in itself – a vast data set from which other GPT bots can learn (as in ‘learning systems”) to spot feigned humanity and thereupon be rewarded for reporting the attempt. (Including college essays or else crime etc.).
Yes, this will lead to an arms race. So? Nations/businesses/NGOs can set up reward incentives to always favor the tattlers.
This is only a completely natural extension of existing systems of tort, transparency, FOI and whistle-blower incentives. After all, when YOU are attacked by one of those terrifyingly smart entities called a lawyer, what do you do about it?
You hire a super smart lawyer of your own.
Indeed, I remain astonished that so few ever mention reciprocal competition among and between AI as the obvious solution, since reciprocal competitive accountability is exactly how we got the miraculous positive sum systems we have, so far.
2. “Provenance is more important now than ever before. User accounts must be more strenuously validated”
Well, yes. The trick is to find ways to verify user accountability without eliminating (as is happening with ‘social credit’ in a rising eastern power) all the benefits of anonymity. We have discussed elsewhere methods by which commercial pseudonymity services might hit a sweet spot, allowing most sites to ban anonymous trolls and ensure behavior accountability while preserving some personal space to ‘be a dog, online.”
3. Some of this is in an extended riff I offered a while back on prospects for a soft landing, re AI. You can find it here: How Might Artificial Intelligence Come About?
== Competitive arenas ==
Yes competition is the c-word that is snubbed by liberals as distasteful… and that is utterly betrayed by the mutant thing that is modern conservatism.
The most creative force, competition is how Nature made us. It is also – in Nature – spectacularly wasteful, inefficient, error-prone and drenched in pain and blood!
The whole notion of our Enlightenment’s five ‘competition arenas’ has been to improve on nature! To set up arenas of ritual combat – markets, democracy, science, courts and sports – where regulated competition can deliver POSITIVE SUM results, the fecund productivity of outcomes…
…while minimizing both cheating and blood on the floor… systems that don’t punish failed competitors with death, but allow them to keep coming back with altered products (if lessened credibility) based on what they had learned.
Alas, in all five major arenas, human nature sows dragons’ teeth that erupt with cheaters, often those who won the last round. Hence regulations… and the need to revise those rules as cheaters adapt… and how we need the internet to learn from those techniques so that accountability systems will kill lies and give us a sixth, useful arena of creative competition.
(And by now is even ONE human still reading this ‘tl;dr’? Never mind, the AIs will have the patience, even if you don’t. But then, if you just read that, then… are YOU an AI?)
Never mind. Here’s some material I crafted in ancient times, maybe two months ago. How archaic and crude!
== The end of photographic reality proof? ==
“On four services alone—Midjourney, Stable Diffusion, Artbreeder, and DALL-E—humans working with AIs now co-create more than 20 million images every day. With a paintbrush in hand, artificial intelligence has become an engine of wow.”
Another disturbingly insightful and eventually optimistic look at the near future from Kevin Kelly. “Because these surprise-generating AIs have learned their art from billions of pictures made by humans, their output hovers around what we expect pictures to look like. But because they are an alien AI, fundamentally mysterious even to their creators, they restructure the new pictures in a way no human is likely to think of…”
Stable Diffusion is the open source system that can be downloaded free and is used already by many for Co-Creative Art Therapy.
Attached are two images my own first experiment. Don’t mind the low resolution. It is, after all, a free version overwhelmed by eager beaver onlookers. (Try inserting THAT phrase somewhere!) I entered: “Odysseus is building his giant wooden horse for Troy, when Deanna Troi beams down from the Enterprise.
Your own favorite prompts are welcome in comments.
Great jeremiads vs bad-predatory destructive trends. A bit short on workable solutions. But sometimes we’re just specialists, y’know?