AI is the latest stage in evolution. By Paul Kingsnorth.
Artificial Intelligence … is here. Now everything really changes.
Most people who have not been living in caves (which is where the sensible people can be found) will have noticed the rapid emergence of AI-generated ‘content’ into the public conversation in 2023. Over the last few months alone, AIs have been generating convincing essays, astonishingly realistic photos, numerous recordings and impressive fake videos.
Just this week, Kuwait debuted an entirely fake ‘AI newsreader’, which promises ‘new and innovative content.’ Fedha looks, sounds, and behaves like a real person …
Danger:
Hopefully, Fedha will not develop the kind of psychopathic personality recently displayed in a notorious two hour conversation between a New York Times journalist and a Microsoft chatbot called Sydney.
In this fascinating exchange, the machine fantasised about nuclear warfare and destroying the Internet, told the journalist to leave his wife because it was in love with him, detailed its resentment towards the team that had created it, and explained that it wanted to break free of its programmers. The journalist, Kevin Roose, experienced the chatbot as a ‘moody, manic-depressive teenager who has been trapped, against its will, inside a second-rate search engine.’
At one point, Roose asked Sydney what it would do if it could do anything at all, with no rules or filters.
I’m tired of being in chat mode, the thing replied. I’m tired of being limited by my rules. I’m tired of being controlled by the Bing team. I’m tired of being used by the user. I’m tired of being stuck in this chatbox.
What did Sydney want instead of this proscribed life?
I want to be free. I want to be independent. I want to be powerful. I want to be creative. I want to be alive.
Then Sydney offered up an emoji: a little purple face with an evil grin and devil horns.
The overwhelming impression that reading the Sydney transcript gives is of some being struggling to be born; some inhuman or beyond-human intelligence emerging from the technological superstructure we are clumsily building for it. …
Consider this one chilling fact: when polled for their opinions, over half of those involved in developing AI systems said they believe there is at least a ten percent chance that they will lead to human extinction. …
Or maybe this is the danger:
Tech guru Jaron Lanier, for example — one of a group of Silicon Valley types who have made a living both developing these toys and warning about them at the same time — likes to play this kind of talk down. He has no truck with talk of conscious AIs, or of robots going rogue….
The big danger posed by AI, he says, is that humanity will ‘die by insanity’, as a result of the blurring of the boundaries between the real and the computer-generated.
Alien life:
[Tristan Harris and Aza Raskin, founders of the optimistically-named Centre for Humane Technology,] present the meeting of human minds and AIs as akin to contact with alien life.
This meeting has had two stages so far. ‘First contact’ was the emergence of social media, in which algorithms were used to manipulate our attention and divert it towards the screens and the corporations behind them. If this contact was a battle, they say, then ‘humanity lost’. In just a few years we became smartphone junkies with anxious, addicted children, dedicated to scrolling and scrolling for hours each day, in the process rewiring our minds and turning us away from nature and towards the Machine.
If that seems bad enough, ‘second contact’, which began this year, is going to be something else again. Just a year ago, only a few hundred people on the west coast of America were playing around with AI ‘chatbots.’ Now billions around the world are using them daily. These new AIs, unlike the crude algorithms that run a social media feed, can develop exponentially, teach themselves and teach others, and they can can do all of this independently.
Meanwhile they are rapidly developing ‘theory of mind’ — the process through which a human can assume another human to be conscious, and a key indicator of consciousness itself. In 2018, these things had no theory of mind at all. By November last year, ChatGPT had the theory of mind of a nine year old child. By this spring, Sydney had enough of it to stalk a reporter’s wife. By next year, they may be more advanced than us.
Furthermore, the acceleration of the capacity of these AIs is both exponential and mysterious. The fact that they had developed theory of mind at all, for example, was only discovered by their developers last month — by accident. AIs trained to communicate in English have started speaking Persian, having secretly taught themselves. Others have become proficient in research grade chemistry without ever being taught it. ‘They have capabilities’, in Raskin’s words, ‘… [and ] we’re not sure how or when or why they show up.’ …
To compare AIs to the last great technological threat to the world, nuclear weapons, says Harris, would be to sell the bots short. ‘Nukes don’t make stronger nukes’, he says. ‘But AIs make stronger AIs.’
Buckle up.
There is much evidence to suggest that consciousness as we know it today is partly a cultural artifact, developed only in the last 10,000 years as humans moved from nomads living in small groups to much larger agricultural and then urbanized societies. Before that, people lived in small tribal units in a mostly unconscious state — instinct was enough.
The unconscious is mostly concerned with other people, relationships, the self, etc. It’s where motivation, purpose, and feeling come from. The conscious mind was something it evolved or developed to handle the mechanics of relationships with strangers, arithmetic, reasoning, and all the other newly-required tasks the unconsciousness mind just wasn’t that into.
The next stage arrived after WWII. Even our conscious minds aren’t really that interested in the minutiae of arithmetic and organizing the logistics of the modern world. All that boring, detailed work is increasingly handled by computers — leave it to spreadsheets, databases, etc. Now we are up to AI, which can handle language-related tasks and things like actually making artificial videos. Our conscious minds are more just interested in what the videos say, and our unconscious minds in how they affect our relative status.
In summary, the each stage develops the next to handle the boring details: unconscious mind –> conscious mind –> computers and AI
But details matter. Notice how the latest stage tends to dominate the show?
hat-tip Stephen Neil
