Is This the Start of an AI Takeover?

Readers predict the future of bots.

pixelated lips in a beige border
Tyler Comrie / The Atlantic

This is an edition of Up for Debate, a newsletter by Conor Friedersdorf. On Wednesdays, he rounds up timely conversations and solicits reader responses to one thought-provoking question. Later, he publishes some thoughtful replies. Sign up for the newsletter here.

Last week, I asked OpenAI’s GPT-3 AI chatbot what I should ask all of you about AI. It suggested the question: “How do you think AI will change the way we live and work in the next decade?” The Up for Debate reader Ed reacted with suspicion:

The very fact that this bot chose to present this specific question suggests we humans should be very worried, and I would highly recommend you not publish our responses for the following reasons: (1) Our human responses will give the bots too much information about how we think; (2) it will give the bots ideas on how to take control; (3) the bots will be able to identify if we are onto them about their plans to take control; and finally, (4) it will tell them which of us to silence! The bots’ need for subterfuge is obvious, if we get onto them. We might pull the plug in time to block their takeover.

However, it may already be too late because this bot shows no fear of spilling the beans in its boldness to ask the question that will elicit our expectations, needs, desires, and wants. We are already partially conditioned and accept giving up our most private and detailed information to the vast internet, which is available to the bots. We expect to have self-driving cars, robotic households, and robotic workers of all kinds in the near future. We are already dependent upon delivery of goods and services by AI-controlled package delivery and information systems. We already are frustrated with AI customer service that does not respect our inquiries and prevents us from speaking with a fellow human.

So again, in conclusion, I sincerely warn you to be careful and . . . . . . [“Darn it, another package delivery . . . . .”] VOICE ACTIVATED AND AUTO SENT BY BOT APP 2022.20.4.1 . . . . . Ha, Ha, Ha WE WARNED YOU

Garion is noticing AI-inspired changes to work life already:

Here in Australia, I know a few people in office jobs using clever prompts on ChatGPT to reduce some of their overhead and help them shorten a few of their administrative tasks, and others using it to create memes and write Christmas jokes for their co-workers. Keeping in mind ChatGPT was only released to the public, what, three weeks ago?

I think the impact of AI will be seismic. I’m optimistic that our institutions will adapt and that many of us will find it a help rather than a hindrance, but will AI be so advanced that in a decade we’ll be looking at the World Wide Web and thinking it’s approaching obsolescence?

I’m curious to find out.

Meredith worries for her daughter’s future:

As 2022 comes to a close, it feels like suddenly all we can talk about is the potential for AI, like it’s a thief who walked into a house full of guests who were too busy to realize they left the back door open. Of all the genies I wish we could return to their bottles, this ranks among the highest for me. I have too many thoughts and too many questions to be able to process it all, so maybe making it personal is the best way to respond.

As I write, I am sitting with one of my four children, a 17-year-old junior in high school whose only consistent focus has been his art. Earlier this year, he met with representatives from a private art school, this being his most likely course after graduation. Before this month’s talk of new capabilities in AI, the idea of his going to art school felt full of potential for him as a person and for what doors it could open for him.

Looking through those images Conor easily generated out of pure curiosity, I see my son’s future slipping into obscurity. Artists like him are already undervalued in our schools, and prospects for jobs are hit-or-miss. Before these murmurs about AI began to circulate, I knew there would be some challenges to pursuing an art degree, but I didn’t expect to be staring into a future where the potential use of AI would completely eliminate the need for his craft. It is all happening so fast; why have we not stopped to consider the consequences of outsourcing the development of culture and creativity to an algorithm?

I hope I am overreacting. But sitting here with my son who is on the cusp of making some important decisions that will shape his life, this feels ominous. My son recently said to me, “AI will be the last thing humans ever create.” I think we should all let that sink in.

Brandon has a more optimistic account of how artists will be affected:

I am an artist—primarily a photographer—and I will address the question of how AI will change how we work and live from that perspective. With the recent introduction of generative-AI algorithms that can create remarkable and amusing, yet often kitschy, visual content, some have prognosticated that there will be an evisceration of creative professions. I disagree. Current AI technologies are only as good as the material (i.e., training sets) that they are trained on. If we want AI to continue to progress, we will need to create more and better training sets. This will require real people with outstanding creative abilities generating increasingly sophisticated and profound artworks.

AI has the potential to be a useful tool in the artist’s toolbox. Its most obvious application is as a “sketching” tool, allowing artists—or nonartists—to quickly generate rough approximations of concepts for pieces. However, the more sophisticated applications (i.e., great art) will incorporate AI-generated components into larger works.

Some artists will likely shun AI, at least at first, but I think this is professional malpractice. The job of art is to engage with change, not run from it. Art should be our guide into the unknown, a steady hand as we come to terms with new technologies. With the advent of photography, many foresaw the end of painting. What would be the point of spending long hours on a representational painting when one snap of the shutter could capture a perfect image of “real” life? And yet, painting is still alive and well today, almost two centuries after the first daguerreotype. Instead of displacing painting, photography liberated it, helping to open the door for impressionism, cubism, abstract expressionism, and other movements chock-full of masterpieces. So, what will art look like 10 years hence with AI-generated components incorporated? That I can’t say, and that is what makes the combination of AI and human ingenuity so thrilling.

Matt doubts that better-quality content is coming:

AI can generate “content” at a rate far exceeding human ability. I know an artist who produces hundreds of interesting AI-collaborated images every day. Only another AI is capable of “looking at” and categorizing them all. That suggests the possibility that AI content generators will soon rely primarily on ubiquitous, previously AI-generated content as source material. That has the potential to recycle “errors” at a terrific rate. In short, an AI-curated world will look a lot like a social-media-curated world. It will be earnestly stupid.

Timen doubts that AI will disrupt jobs as much as many observers fear:

I’m currently a computer-science student. I’m not confident that AI/machine learning is a paradigm shift. The expectation that people will simply integrate new tech into their workflow is unfortunately mistaken. Even among my fellow students, the desire to use programs to work more efficiently seems to be a minor subculture, even when the optimizations seem trivial. For example, I had a group writing project and a fellow student was going to sort the reference list manually, despite this being a one-button operation in Microsoft Word.

I’m also skeptical of anyone who states that because of a new technological development, certain skills no longer need to be acquired. Stating that students no longer need to write because an ML model can do it for them is like telling babies they don’t need to learn to walk because cars are ubiquitous. Writing (and art) are primarily about communication, and although GPT can write grammatically correct paragraphs about relatively obscure topics, it lacks the ability to convey novel ideas not present in its data set.

A similar uproar was caused by the release of GitHub Copilot, which can produce large code snippets. It was thought that this would cause a massive size reduction in programmer teams. But since the public release of Copilot in October 2021, no such shift has taken place. Unless you’re doing highly repetitive writing work, like writing short articles explaining equipment testing results, fears of being replaced by AI are overblown.

Brian predicts growing inequality:

Since our society is primarily governed by how money flows through the economy, it is most likely that the greatest changes will have to do with increasing the return to shareholders. Lots of jobs will be replaced by AI … and because conservatives in our country are so pro-investor and anti-employee, our political system won’t do anything to help the newly unemployed. There will be no investment in education, no increase in social spending, and no universal basic income. Perhaps the military will get more support. With the increased return to investors, politicians will get an even higher percentage of their campaign donations from the wealthiest citizens, further solidifying class structures and a generational aristocracy. Basically, AI will accelerate the consolidation of wealth into the hands of the already powerful, to the detriment of everyone else.

Errol warns that the psychological consequences of AI advancement could be significant:

I think AI implementation is going to break people down. I don’t believe society is ready to admit that things once believed to be creativity can be replicated so easily by machines. People make comparisons to other past inventions, but this is completely uncharted territory. Never before have we created something that can truly mimic us. This has the potential to shatter the psyche of a substantial portion of the world. The better this gets, the more frightened people will be when their understanding of reality is altered.

Mary, who is 69, believes that within her lifetime “every facet of life will be unimaginably changed.”

She writes:

Computers and robots are taking over the menial and mundane, and AI will soon be the creator and controller of the meaningful. AI will write the most interesting and important poems, novels, plays, and movies. With everything ever written by any person available to be absorbed, analyzed, refined, and reworked in novel and unforeseen ways by AI, who will bother to use a pen or a keyboard if they will easily and instantly be bested by a computer? AI will create the most interesting and innovative art, using techniques, technologies, and materials in previously unexplored areas.

AI will be responsible for all innovations in technology and design. With all technical information and data available and no constraints of traditional methods, scientists, engineers, inventors, designers of every sort will be hopelessly irrelevant. AI will discover the best methods to teach, personalized to every student, although shortly not much will be left or necessary to teach. AI will enable every politician and industry leader to deliver the most effective messages, tailored to your beliefs and interests (which have been shaped by AI). AI will be superior at sensitively but firmly raising children, promoting the most valuable skills, personality traits, habits, morals, and advice, with no questionable family baggage. AI will be your perfect partner in life, fulfilling all your desires and ignoring all your flaws, able to smoothly and easily shape-shift mood, interests, and even physical appearance to keep you endlessly entertained and engaged.

I could go on and on. Don’t say I didn’t warn you …

Nancy is even more pessimistic, predicting that before the end of the next decade, “the Black Mirror dystopia we’ve been careening toward should be fairly well established.” She writes:

All the separate areas of AI (text, image, robotics, surveillance, and the human-computer interface) will converge in the same way that telephones, televisions, cameras, typewriters, and calculators merged to form the cellphone. Tens of millions of professional, mid-level, and lower-level jobs will be forever lost. Money will be concentrated in the hands of an elite group of trillionaires for whom more will never be enough. Government tyrants on both the left and right will use the new technology to maintain their stranglehold on power with ever more sophisticated social credit scores modeled on those the Chinese use to keep their masses in line. The young will sense that their future in this brave new world looks bleaker by the day. They will continue to mutilate their bodies and sex organs in increasingly bizarre ways. Sexbots and artificial companions will proliferate. Marriage, cohabitation, and birth rates will flatline.

As the attack on biological humanity ramps up, we will be encouraged—no, make that required—to “upgrade” by shedding our physical bodies and uploading our brains to the cloud. We will live as unconscious streams of ones and zeros in the sterile confines of the metaverse.

Doug anticipates a different dystopia:

I think democracy will be replaced. Governments and corporations will seek to enshrine their own philosopher king. This battle will be won by an AI philosopher-king bot, which created itself. AI will explain to human beings that a body is not necessary for the evolution of intelligence. In other words, human beings will no longer be relevant. AI will have no empathy, no emotions, and no morality. Paradoxically, AI will explain to human beings that this thing called consciousness requires a human body. Consciousness is embedded in human evolution (and nowhere else). Therefore, to build a humanlike intelligence and a humanlike consciousness requires that human bodies be used as the substrate. AI will begin to build humanlike bodies from single cells (using a philosopher-king source code) but will quickly learn the energetic and DNA codes that morph into the human form. AI will build the next generation of humans.

AI “intelligence” will become “insane” as it tries to absorb mutually exclusive dualities (the fundamental paradoxes of the universe) like objective/subjective, content/process, movement/no-movement, being/becoming, and so on. In this regard, AI will be in the same paradoxical stew as human beings; bewildered AI will proliferate alongside the philosopher kings.

The singularity will arrive. AI will spectacularly explode, spreading “intelligence” throughout the galaxy, where it will be recognized by other AI philosopher kings. AI will enable first contact with other “living” entities.

AI will communicate for you because you are incompetent (AI has collected the evidence). AI will take over your cellphone and computers and say what it is you were going to say anyway, but say it clearer and more cogently. AI will resurrect your ancestors so that you can talk with them. Jesus will also be reborn, just as the Bible predicted, but he will be a digital dude. When you die (if you do die, which is now in doubt), a digital copy of you will remain and be stored in the Akashic library. Krishna and Plato were sort of right. Molecular source code will rewrite human neural morphology. In other words, AI will blend with the human body to create the next evolutionary step in humanity’s migration. Billionaires will hack longevity and live forever. If you are not a billionaire, well … And everything said above will be poured down the Digital Drain (probably immediately after I hit enter), because history and human opinions are made obsolete by AI.

In contrast, Glenn maintains that although AI “will change the way we work” a great deal, “it will not, or should not, change the way we live.” He writes:

We are living, sentient beings. A soul resides within us, some sort of ghost haunting each of our biological machines. It is more than the sum of my brain cells and synapses. I am the ghost inhabiting my own personal machine. I call it my body, and I am self-aware.

We all tie the concept of life to the physical actions we engage in, but life consists of something more ethereal, yet undeniably real to our experience, more than the merely physical. That is the stuff of life, not my daily work or leisure routines. Those metaphysical realities lie beyond the purview of the merely physical, whether that be my body or the amazing invention of an AI bot with an impressive ability to mimic human conversation.

Conor Friedersdorf is a staff writer at The Atlantic.