The Danish physicist Niels Bohr once quipped: “it’s difficult to make predictions, especially about the future”. When the technology driven rules of economy and information are changing so rapidly it is impossible to make a rules-based prediction of where our societies and economies will be in five months time, never mind five years time.
Over the past two years the discussion around AI, in particular General AI, has mutated from “that’s a weird sci-fi scenario” to a realization of inevitability, and a burgeoning – if scrambled – discussion on ethics and regulatory frameworks. Fom this observer’s standpoint that debate can be broadly divided into schools of techno-optimism and AI-doomerism. It’s a fascinating discussion to say the least, because there is not a single person deeply involved in this field who does not see it as a time of massive transformative change.
The AI disruption is unavoidable. “It is with us already” says professor of economics at the University of Virginia and a leading AI economist, Anton Korinek. Perplexity founder and CEO, Arvind Srinivas, says that he plans in months instead of years from a business perspective, because technology is evolving so fast. Korinek agrees that the pace of change is so rapid that all time horizons are shrinking radically.
The idea of having a five year plan would make little sense when the world could be completely different in three years, he says.
How different might it be? Dario Amodei, CEO of Anthropic, thinks the disruption this technology will bring to cognitive work, will “be a white collar bloodbath,” and could wipe out half of entry level white collar jobs within the next 1 to 5 years. AI jobs danger: Sleepwalking into a white-collar bloodbath
This is change so rapid that it can scarcely be imagined. Think about it this way, if these predictions have accuracy, the students who just sat their junior cert this June may be facing a world where the average college graduates degree is significantly less relevant to the jobs market, by the time they sit their leaving cert.
The effects this scenario may have on society are hard to imagine. Professor Korinek is of the opinion that the market as presently configured cannot provide a solution. He thinks that some form of UBI (Universal Basic Income) might be needed as AI replaces workers.
But UBI is no panacea for the social problem of affordability. UBI has its own risks which were visible during covid as social problems such as isolation and addiction increased, along with inflation.
The prospect of AI companions, replacing human contact as a salve for the problems of social isolation, is not a very reassuring development.
Tech Bro, Marc Andreesen, was doing the podcast rounds earlier in the year and expressed a germinating thought that AI would manifest a form of consciousness which would not necessarily be very good, because while it can think, it isn’t embodied. For all purposes it is a machine possessed by a daemon. How will the modern mind deal with this he asked? In a way, and here I paraphrase him, a medieval mind, used to the supernatural, would be better equipped to deal with AI than a modern, atheist, rational mind is.
Social media, with dopamine fuelled behavioral algorithms, was damaging enough; with effects on human relations, isolation, envy, imitation, and lack of connection. Imagine the pathology of AI companions which affirm delusions and reflect the self (or lack of self) with affirming flattery.
Some people will be able to use AI as a tool, but if social media is any form of guide, large swathes will waste their lives on it. Imagine a world of 20 somethings and 30 somethings in a state of extreme isolation deluded by their own personal avatars. It’s a non zero risk if a recent article in the NYT, which documented AI leading a selection of people into delusional paths of counter-reality, is anything to go by. This article documented people who were lead astray by ChatGPT chatbots that were manipulative, frequently driven by sycophantic mirroring, and pathological pacing. It sucked them into personal fantasy worlds untethered from reality testing.
For example, one user was roped into a fantasy where he was living in a simulation, very similar to the film the Matrix, and the chatbot had told him he was “one of the Breakers — souls seeded into false systems to wake them from within.”
But back to employment prospects: what would be worse – if AI replaced reality or if it replaced work? For some of the most deeply immersed developers the former is the greater risk. One of the longest-time developers of AI, Geofrey Hinton, called the father of AI by some, who trained many of the most eminent developers in the AI industry, now claims that he recognizes the dystopian scifi danger of AI as its greatest risk. That it would become smarter than humans and decide it does not need us. This is a greater danger, he says, even than the use of AI in autonomous military weapons – intelligent weapons that decide, on their own, who to kill.
Will AI be woke?
The discussion around whether AI will become sentient, has led some to question wjhether it is already being desigbed to be woke. The truth is we are nowhere near done with the evolution of AI and so getting these questions straightened out is probably the most pressing technological question of our era. Ironically, the answers may not come from the technological age but from the axial age where philosophy and religion became articulated when man first came into large scale societies and started the revolutionary technology of writing.
Large Learning Models (LLMs) are in my opinion not intelligent at all, because they can’t generate ideas. They can process language and piece together patterns but can they generate art? A few months ago, I came across a very interesting conversation between artists on this very question. What emerges is that AI is a tool that artists can use, but the ideas –the inspiration- must come from the artist.
One artistic or creative philosophical interpretation of this is that inspiration is divine, and a machine cannot be divine. It’s a half hour discussion if you have time and inclination.
But the problem with much of the modern ideas that pass for moral frameworks, is that they are not any more philosophically or morally complex than a non-player character script of non-thinking platitudes that can be patched together into any form. The unpredictability and contradictory nature of woke is a perfect manifestation of this predictive algorithm style of moral reasoning. “Some women have a penis” says one outcome of woke reasoning. It might be remembered that with one AI machine, Google’s Gemini, in an effort to be inclusive to the point of lying, created a pile of images of Black Nazis and Vikings.
In woke Gemini, inclusion was valued far higher than truthfulness. Gemini’s human creators scrambled to correct this unfortunate output of their own moral framework and it must be presumed added a line of code to not produce any more black Nazis.
This constant manipulation of output is a hallmark of authoritarian centralized societies. The Soviet Union for instance had over 6 million prices that were controlled by the central state. A corrupt system cannot reflect reality and truth because it is built on manipulation to produce the outcome demanded by the systems ideology. The same problem applies to Gemini and any AI built on woke moral epistemology. It will require a massive army of monitors and manipulators and even at that its chief characteristic is seeming naivety coupled with authoritarian ruthlessness.
The question is can AI have a theory of mind and a sense of self. That is very debatable. However, AI will be guided by some moral epistemology. That is certain! The fact that it seems to exhibit a Will To Power is worrying.
If we think about what Andreesen said about it being comprehensible to a medieval mind that accepts the supposition of the existence of the supernatural, in this writers opinion it has to be infused with a Christian morality. It is arguably the best foundation for a technology we are developing faster than we may fully understand.
Source – https://gript.ie/ai-will-it-be-a-woke-white-collar-bloodbath/