Four possible futures for AI and society
Scenarios after Jim Dator
How might AI change society?
Right now I’ve been writing to that question as a chapter in my new book. Peak Higher Education is an exploration of one of my scenarios, wherein American higher education gradually declines in many measures. In the book’s middle I raise three giant topics, teasing out how academia might respond: climate change, emerging attitudes towards the future, and AI. In that AI chapter I start with big picture issues, including the emerging industry’s fragility, then I offer four scenarios for how society might change.
Today I’ll share those AI scenarios as an extract/draft. I think they might be of use and would love to hear your thoughts as I develop them further.
James Dator (University of Hawaii) first published his four-fold model back in 1979, and has adumbrated it since, as have other futurists. These are templates or images, archetypes or generic stories of how big changes might occur and how we can start to anticipate them. To summarize, each is a verb pointed at the macro level:
Grow: this means social growth, which often includes economic expansion, increased energy, more and improved technologies, longer lifespans, and so on. This can also describe related situations where growth is the theme, like a growing nation or enterprise.
Collapse: a society which suffers a reduction along various metrics, even to the point of extinction. Quality of life declines along the way.
Discipline: a society which (generally) organizes under a (generally) shared ideal. Dator gives the example of anticonsumerism, or a shift to a conservation society.
Transform: society becomes stranger, with new ideas and practices appearing, new problems resulting, and changes to life keep coming.
I have always applied these to forecasting on multiple levels, from the macro to the micro.
Recently I’ve been thinking of how this foursome might suggest routes for AI-driven social change. Initially I applied Dator to AI itself, imagining ways it might grow, collapse, become disciplined, or transform, but those seemed too simple and already discussed. Instead I’d like to consider how society might change as it apprehends AI and as the technology transforms it. By “society” I want to be a little vague to start with, allowing for given nations and the whole world.
Consider these short scenarios. I’ve added two versions of one:
Grow AI helps drive social expansion. The economy improves as productivity rises with the help of AI efficiencies and new, AI-backed businesses expand and become richer. Health care, transportation, governmental administration, and other fields become more productive and effective. More people have more access to more services, from education to law, health care to professional development. Gross domestic product (GDP) rises.
This is net growth. Similar to the industrial revolutions, every AI advance caused some businesses to shrink or fail, yielding substantial unemployment. Yet the growth in jobs was greater, with new professions emerging to help make the AI revolution work, leading to an overall improvement in the “creative destruction” pattern.
Culture experiences something of a Renaissance as AI enables more production and new forms. Gaming, video, audio, images are grow, especially as new talent (who could only do this work with AI help) floods in.
AI’s impact on the climate crisis has become a net positive. While training LLMs requires a great deal of energy, it’s less than it once was, thanks to efficiency improvements. Meanwhile, climate scientists, government officials, corporations and others use AI to improve climate mitigation and adaptation strategies and operations.
Education has mastered AI as it did earlier technologies, such as calculators and the world wide web.
Collapse v 1 Generative AI collapses like the Metaverse or Second Life, toppling from a peak of investment and excited down to a niche. A combination of forces make this happen. Judges in copyright cases rule against one LLM provider, then another. Investment starts to drop. Public attitudes become more skeptical and active resistance grows. Businesses reduce their AI offerings. OpenAI goes broke. Generative AI still exists, but only for specialists on the margins.
Social impacts spread. The loss of billions of dollars in investment ruins some careers, but more importantly causes problems for other investments, leading to more projects beyond AI falling apart. AI professionals suffer widespread unemployment. De-AI-ing government and business becomes a painful chore, which some AI gurus grimly work at to pay their bills. Popular attitudes against Silicon Valley sour, with users dropping away and some major firms losing value.
The Chinese government punishes AI leaders in various ways, mostly through career endings. The downfall of AI further injures that nation’s troubled economy, which might tip things towards social unrest.
Activists who saw generative AI as a climate threat, flush with success, turn their attention to other enterprises they see as worsening global warming: airlines and fossil fuel companies.
Educators generally stop teaching with or about AI, outside of computer science classes.
Collapse v 2. The implementation of AI at scale causes many social problems, beginning with the degradation of the labor market. Some current workers find their hours reduced to part time and/or their work deskilled as AI takes up more functions. Others are laid off as AI largely replaced them, albeit with some humans to wrangle and check for quality. Management and owners of AI firms see their value soar as their companies do, while actual AI hiring is paltry, further exacerbating income and wealth inequality. Outside of AI companies, many business struggle with losses as they fall behind in an automation arms race or shut down when AI-enhanced firms outcompete them. GDP fluctuates, not yielding an overall growth curve.
Governments struggle with the crisis, trying to maintain social services increasingly in demand while tax revenues run low. Politicians are split, with some opposing AI, urging higher taxes on AI companies and the rich, while others seek the support of those same actors. Political campaigns and government communication regularly use AI-generated content.
Cultural production suffers as studios lay off talent and replace them with AI teams. Audiences split on the results, leading to a decrease in purchases and audience share. It becomes a popular diversion to spot content flaws in games, movies, podcasts, and books.
Overall energy demands for AI functions grow, even as new configurations reduce their electricity and water demands, because the requirements for quality output demand more and more computational runs and AI usage keeps rising. Popular antipathy towards greenhouse gas emitters settles on AI firms, leading to attacks on datacenters. Researchers make the case that AI is worsening the climate emergency.
Education sees more teaching about AI, largely as a way to give students a better chance in a worsening career market.
Discipline Society splits in two, based on considered attitudes towards AI. AI users and supporters gradually build out services and implementation throughout the world, evolving protocols and practices. AI opponents carefully carve out “no parrot” and “Clippy free” zones and lifestyles. These attitudes become ideologies which spread throughout society, from politics and economics to culture and relationships. Large administrative systems are reconfigured to adhere to one side or the other.
For years this ideological divide plays out to various degrees of visibility. Government and business services might proudly proclaim their use of AI as an aid to user experience, while creatives boldly offer 100% no-parrot certificates on their wares. Youth subcultures rise, fall, and change based on performative and practice AI attitudes. The rise of AI-enabled robots makes one’s position on the divide even more clear.
In the climate change world the divide is stark. Pro-AI researchers and officials rely on AI to analyze sensor data, to generate models, to create new forms for climate mitigation and adaptation. Anti-AI people cite climate impact as a major reason for their position and actions.
We can determine an individual’s schools AI stance from its staff and marketing. Curricula divide over this, with some academies emphasizing human learning as a form of humanism or classism, while others describe preparing students well for something like a cyborg world.
Transform The rollout of generative AI and its knock-on effects increasing make society stranger as institutions and practices destabilize. An AI-assisted robot renaissance sends bots all over the lifeworld.
New forms of government operations appear, such as digital twins of politicians, artificial characters of officials. AI-generated content increasingly appears through state publications, not without incidents.
The business world is more chaotic as new customer service and workplace practices appear, mutate, or disappear. New businesses and types of companies appear as AI services proliferate. Productivity climbs upwards, yet a big churn exists with business start ups and closures.
Personal relationships with AI, then with AI-enhanced robotics, proliferate while remaining controversial. Lonely people from children to retirees spend more time with automation and less with other humans, especially in person; cohabitation occurs and the humans demand privacy (“It’s nobody’s business what my bot and I get up to!”). Complaints about humans “logging out of society” rise, leading to bot-shaming. Less intensively, people use AI characters as avatars or fronts for digital situations, then for offline ones, with avatars answering doors and telepresence robots rolling into offices. Various bots serve religious needs, from advice to spiritual leadership.
AI-generated cultural works proliferate and become stranger, a new wave of surrealism. New kinds of art appear in movies, body modifications, architecture, and comics. AI-assisted neural translation of human dreams, the thoughts of locked-in people, and of animals begin to change our conception of our selves.
AI-generated climate action is occurring. New designs for wind turbines, solar panel sites, buildings, and geoengineering appear.
That’s where the set stands for now. Each contains elements quite different from the rest, but there are some resonances between them at other points.
What do you make of them? Which seems most likely to occur? How might your own world change when living in one?
For me, back to writing…
Transform resonates with me. Society already has a firm foothold here from hedonism. As weird as the last decade has been, an even weirder future would not be a surprise.
Very interesting Bryan.
Ever since reading New York 2140 I’ve been making notes on a project I call Kisangani 2150. The idea is to take the world Robinson created in NY2140, run it ahead 10 years, and focus on Kisangani. Kisangani is in the Congo at the place where Conrad’s Inner Station was located. Basically, Stanleyville.
I’m imagining a very specific connection between NY2140 and K2150. As you recall, one of the settings for NY2140 was a nightclub called Mezzrow’s. Though I don’t think Robinson explained the significance of that name, I assume he’s alluding to Mezz Mezzrow, a minor figure in the mid-20th century jazz scene best known for supplying marijuana to the musicians. He was also a musician and helped put together recording sessions. Anyhow at the very end of the book some of our heroes go to Mezz’s and dance to the tightest West African music anyone ever heard. Well, the world of NY2140 couldn’t have functioned if there weren't a lot of clubs like that (just as there are in our world). They provide essential social glue and breathing room. It’s just that they were peripheral to Robinson. Not so in K2150.
The way it looks to me, the world of K2150 combines two of your scenarios: Discipline and Transformation. The New York world is already deep in the Discipline regime. All those supertall buildings and the floating villages, they’re the AI adopters. The Anti-AI people inhabit the liminal zones (I forget what Robinson called them).
But Transformation is underway in the Kisangani world. But they’ve adopted a different style of AI tech. Not the resource-intensive tech of the New York world, which is an extension of the current scale-is-all regime. While scale doesn’t turn out to be all, it does quite a bit, enough that it dominates. The Transformers, however, go another way, in part because they can’t afford the BIG TECH, but also they’re not wedded to that worldview. They a adopt a more organic approach (I can probably dream up technical details, but not here; you can find some of them here: https://independentresearcher.academia.edu/BillBenzon/Cognitive-Science-and-Psych). Kisangani emerges as a receiving place for fugitive “organics” from all over, including China, Japan, and India.
Of course, there’s nothing predictive about this. It’s just something I’ve been thinking about. I may use some version of it for the final chapter of a book I’m planning with a tentative title: Welcome to the Fourth Arena, which is an amplification of this article: https://3quarksdaily.com/3quarksdaily/2022/06/welcome-to-the-fourth-arena-the-world-is-gifted.html.