“While AI is exploding now, it is not happening in isolation. Other factors are having a powerful impact on individuals and social systems, namely:
- Climate change
- The global housing shortage and refugee crisis
- Changes in attitudes toward work and economic survival following COVID
- A global rise of fascism and authoritarianism in the face of staggering economic inequalities
“Some would set AI advancements and technological development apart from these factors. I would not. Rapid technological developments are still largely subsidized by high-net-worth individuals through VC investing, tech incubators and the like. No one expects AI to be immediately profitable. But, should investor sentiment change another ‘AI winter’ could appear as quickly as investors lost faith in banner ad click-through rates in 2001.
“What I can predict for 2040 remains contingent on the unpredictable nature of these issues. Some might argue that AI tools will go to work on problems of atmospheric carbon capture or refugee distribution, with potential solutions within reach as surely as AI is driving very real medical advancements in chemistry and genetics. This is possible, but assuming AI can untangle our fossil fuel and climate dilemmas amounts to blind faith in AI’s goodness as much as the irrational fear of Skynet amounts to blind faith in its badness.
“AI critics and skeptics seem to fall into two camps: the bias-and-danger-right-now camp and the far-future-dark-singularity camp. Both should be taken to heart. We need slower and smarter (and more explainable) AI tools right now, and we need wiser exploration of the far-future implications of current AI infrastructures, patterns and governance.
“I hope wiser exploration of the far future of 2040 can come out of this particular study. Work like this should be a springboard to further research, perhaps by a generously-funded global consortium empowered as a governing body. It might be modeled on the World Wide Web Consortium or a more comprehensively binding group in order to also take into account corporate proprietary technology that is resistant to the controls needed to protect the Earth and its living populations.
“To project forward to 2040, let’s assume such a body is created and exists. Let’s assume our tech industry overlords have altruistic motives. After all, they are driven to create benefits and consumable tech for their super-rich funders, if nothing else. Such a body could come from the worlds of Davos or the Aspen Institute to forge a governing alliance between big tech and global financial power.
“The Low-Code/No-Code Internet they might create would be both good and bad. On the upside, it could be like Geocities in the 1990s, but for tools and apps, as the barrier to a more sophisticated and functional web presence falls to near zero. This could be a boon to small businesses, rural economies and community organizing. On the downside, all communication channels are likely to become clogged with frictionless, AI-generated content, scams, deep fakes and snake oil vendors run amok. Perhaps AI search will also become more sophisticated, better able to tell valuable content from noise or harms.
“It seems clear conventional ‘search engines’ will not be up to the job much longer. Their replacement by summarizers and conversational agents (some already passing the Turing Test) is well underway in 2023. Search engines in 2040 will be remembered as artifacts of a quaint interregnum that lasted a mere 25 years. They’ll be in a museum with Archie and Gopher and HyperCard.
“Benign shifts in our Internet lives will matter less in 2040 than they do now because there will be no boundary between online and offline life (presuming civilization has not fully collapsed). What we consider ‘meatspace,’ or our walking-around lives, are what will have changed the most, aided, facilitated or made worse by the speed of exponential AI/ML development (both specialized and general), accelerated climate change and possibly also by a neo-feudalism fostered by decades of uncorrected disparities of wealth.
“Any affordable consumer devices that can be made rechargeable, portable and unconnected to the power grid will be, including all forms of lighting and illumination. Nikola Tesla dreamed of wireless light. It will be a reality. Power outages will not be ‘blackouts.’ Low-power-using, motion-detecting, off-grid LED lighting will be ubiquitous. It will also be so indirect and ambient outdoors as to bring back the starry night sky to cities. And the nature of the power grid itself will have changed by 2040, and not just from AI-driven load balancing and anomaly detection (specialized AI). 2035 is frequently cited as a tipping point for climate change. Given the temperature records set in 2023, many climate scientists are scrambling to revisit their data projections, fearing accelerations and knock-on effects not previously accounted for.
“Assuming more-frequent weather and climate disasters between now and 2040, I expect dependencies on a centralized power grid to change substantially. Extreme weather-related outages will lead to most permanent housing being built with a back-up power source or generator, likely with sophisticated routing to essential systems to moderate the impact of outages. Add to this the proliferation of cheap, rechargeable, non-grid-dependent consumer devices and the ability to feed power back into the grid. It will be a distributed system, in other words, a power grid that works like holiday lights: one goes out, the rest stay lit. I’m referencing all permanent housing for another reason.
“I expect a larger number of people will be living somewhat normalized, nomadic lives, willingly or unwillingly, extrapolating from how little market forces are reacting to the current U.S. housing crisis and how climate disasters will increase the number of unhoused or displaced people. By 2040, this semi-nomadic population could be quite large. It would also be large consumers of off-grid or rechargeable devices. Portability, for them, would be critical.
“This movement could also be driven by changes to the world of work, particularly white-collar work, which is moving out of expensive city office buildings and into a virtual network that could level off into a kind of cottage industry of home workers (at least after the pricey corporate office leases and tax breaks run out). These economic systems are made possible by the accelerated impact of AI/ML workplace tools, while also complicated by exponential climate change effects, which no country in the climate accords seems to have the political will to address.
How does this future look socially? … I believe this will go beyond the wealth polarization seen in the Victorian Age during the Industrial Revolution, for instance, to a kind of neo-feudalism, pricing the best tech out of the reach of the ‘serfs’ in their RVs, tiny homes, shipping container villages, Hoovervilles and converted office building ‘dormitories.’ Even as they abandon contributing to the good of the larger social infrastructure and instead use their extreme wealth to create new kinds of castles and moats, to stand with pre-ghost-visitation Scrooge and send the less privileged to die in the overheated countryside and ‘decrease the surplus population,’ they will lose more than they expect.
“By 2040, I would expect to find a number of climate no-go zones: areas with no ground water access, burnt by industrial waste, with unmoderated deadly heat, perhaps even moonscapes with no vegetation. New deserts will form, just as parts of the Sahara cover what was once a lush landscape. The Amazon basin itself could become a desert. Australia’s inner desert could grow to cover most of the continent. And many hydroelectric power sources, such as the Hoover Dam, could be at risk. Socially? Well, acclaimed author Margaret Atwood imagined what might happen with polarized wealth and technology in her ‘Maddaddam Trilogy.’ Susan Collins, author of the ‘Hunger Games’ series, envisioned it as well, in the contrast between Panem and the Districts.
“Quick mobile egress in a fast-changing world will be as necessary as a fire escape in a building is today, because a flood could come from one direction, a wildfire from another, a hurricane from another and wildfire smoke could envelop the atmosphere, as it did in the northern U.S. this past summer.
“I see two worlds emerging, even in the richer, industrialized spaces, with the wealthy moving through and paying a premium for more secure transportation ‘corridors’ connecting their technologically-sophisticated enclaves. Everyone will either live in an RV or own one, even the very wealthy, who will ensure their relative security of place in compounds with bunkers. Those in the more authentic world, will break from the ethos of accumulating things, of unthinking consumerism, perhaps from having lost their things in weather-related disasters, and instead finds community in mobile groups, parked at sympathetic farms, Walmart parking lots, campground ‘villages’ or spaces designated for refugees.
“How often they have to move will depend on the relative safety of these transformed sites. They are connected and empowered, however, and technological tools facilitate their connections and communities, just as CB radios once connected truckers on the road.
“The merely rich, the super-rich and the billionaires have already begun constructing their bunkers, their compounds. They will have access not only to AI-powered electronic security and private armies, but also the most advanced and expensive AI-driven medical tech. They will be the ultimate audience and consumers of the most advanced machine learning innovation. I believe this will go beyond the wealth polarization seen in the Victorian Age during the Industrial Revolution, for instance, to a kind of neo-feudalism, pricing the best tech out of the reach of the ‘serfs’ in their RVs, tiny homes, shipping container villages, Hoovervilles and converted office building ‘dormitories.’
“After all, wealthy people are the ones who invested in and paid for the tech. They naturally expect to have the first crack at consuming it. But even their fortified compounds and bunkers can’t protect from the full ravages of climate change, the unearthly, smoky orange haze, the fires, the rising water, the severe storms. They will need to be mobile too. I’m sure they expect multiple homes, yachts, helicopter pads and private jets to take care of it. If need be, they’re ready to go to ground. COVID, for them, was a rehearsal.
“The rich will also suffer in less visible ways. Even as they abandon contributing to the good of the larger social infrastructure and instead use their extreme wealth to create new kinds of castles and moats, to stand with pre-ghost-visitation Scrooge and send the less privileged to die in the overheated countryside and ‘decrease the surplus population,’ they will lose more than they expect. Two that are top-of-mind for me:
- “Above all, human innovation will suffer due to the lost potential of those who, if they had lived in more charitable circumstances, might have come up with better solutions for an inhospitable planet than a Malthusian die-off as a bargain, as happened in the prosperity that grew out of the Dickensian 1800s.
- “And valuable data on humanity will be missing. Machine learning, for all its promise, relies on data. That data, fed into a giant hopper to train the dreamed-of ensembles of specialized and general AI models, must necessarily reflect ourselves back to us. While creativity, with surprising analogic connections, turned out to be ridiculously easy for AI tools to master, the ‘mind’ of AI will always be human society’s mirror image. If AI agents become biased and fascist it is because our cultures are biased, with visible and invisible fascist tendencies. AI job applicant screening tools prefer the names and qualifications of homogenized white men who come from money because the data collected gives those qualifications preferred treatment.
“AI/ML tools learn the essence of who we are better than we are able to see in ourselves. We can program the algorithms to ‘remove bias’ from the data at the very risk of destroying the ‘accuracy’ and ‘truth’ of what the data represents. To remove bias intentionally is to ask the algorithm to accept a lie about the source data, the training data, the synthetic data. To make the AI a less accurate mirror of who we really are, warts and all.
“If the presumed Malthusian bunker-dwellers of 2040 cut themselves off from the larger community of humanity – from the ‘surplus population’ – they will not only be poorer for the loss of the minds of the creators who never lived or never found their potential, they will also have much more narrowly-constructed AI tools, because they will have lost the richness gained from more diverse population who could contribute to a more diverse data set to train and create better models.”
This essay was written in November 2023 in reply to the question: Considering likely changes due to the proliferation of AI in individuals’ lives and in social, economic and political systems, how will life have changed by 2040? This and more than 150 additional essay responses are included in the report “The Impact of Artificial Intelligence by 2040.”