Possible Futures of an AI-Driven Society: A Techno-Political Reflection

Definitively, this is not the first time IT professionals have faced a wave of catastrophic hype. In fact, it isn't even the first wave since I started working in the field back in 2014. Anyone who has been in the tech industry for longer has certainly seen this movie replayed multiple times. Back in college, a professor recalled how the "end of the software engineer" was already being declared when the first intelligent IDEs emerged—integrated with compilation and interpretation environments capable of pointing to the exact line of code containing an error. The logic back then seemed bulletproof: "If the software knows where the error is, it knows how to fix it. The next step is learning how to write itself." We know history didn't unfold quite like that. However, ignoring AI's current impact on our profession and society—or praying for the bubble to burst so we can go back to working exactly like we did before Claude Code and Codex—is an exercise in wishful thinking. It's frequently argued that the cost of maintaining current AI infrastructure is unsustainable. Yet, historically, every revolutionary technology is born expensive: either it adapts its format and consumption model, or the underlying hardware and resource industry receives the necessary capital to scale up and support production demand. From this point forward, I’m stepping away from a strictly technical view to offer a personal political-philosophical reflection. Consider this a head-up. A second disclaimer (though already somewhat implied): none of the hypotheses here involve machines ruling over humans, nor scenarios resembling the movie Idiocracy. AI as Preparation for Society's Demographic Shift I like to view the AI race through a different lens: what if one of the primary drivers (whether conscious or not) among billionaires is preparing society for the global population decline we are currently undergoing? You don't need an Economics degree to realize that our current productive model—in terms of product and service availability—cannot sustain itself through a sharp demographic contraction that hits both the workforce and the consumer base. The COVID-19 pandemic offered a preview: there wasn't a massive drop in total global population, but rather a temporary unavailability of labor due to lockdowns. That alone triggered severe shortages and friction in accessing essential goods and services dependent on human labor. Until generative AI proved its ability to replace human effort in cognitive and intellectual tasks (given that physical robotics hasn't seen the same recent leap as software and natural language processing), there was no clear answer for how society would navigate population aging and plummeting birth rates. A vast hospital network exists in major cities only because people are continually born, study, and become doctors, nurses, technicians, and sanitation staff. Likewise, demand for that infrastructure exists only because there are patients needing care. Expecting population decline to shrink symmetrically on both supply and demand sides is unrealistic—structural imbalance is the far more likely outcome. We already see a practical application of AI at its current capability level in customer support. It is increasingly standard to use AI agents as the first tier of interaction, escalating to a human representative only when the AI cannot resolve the issue. While this concept predates generative AI through rigid rule-based chatbots, it has evolved from decision trees and keyword searches into contextual agents capable of querying complex datasets. Extending this trajectory, it's natural to predict that first-level requests across industries will be handled by AI by default: Healthcare Triage: Picture an autonomous room in an emergency department where an AI directly evaluates vital signs for non-critical patients. Technical Operations: A support agent capable of guiding a customer through troubleshooting a router or machine, or executing remote diagnostic and reboot operations directly before walking the user through next steps. In this model, human professionals step in only for complex cases requiring subjective judgment, visual nuance, or deep contextual understanding that AI is unprepared to handle. To an extent, this transition is already underway. Perhaps the time required for AI to handle more complex roles will align with the window in which overall population and demand for high-variety services shrink. In that scenario, the goal is ensuring no one is left without medical care, groceries, or public transit. It's worth noting that this reflection doesn't solve how displaced workers in low-skilled positions would be reallocated—that remains a far broader social challenge. The End of Work Does Not Mean the End of Income Stepping into the realm of the singularity—where AI and robotics assume nearly all productive functions—humanity would theoretically be sustained by the value generated by this digital workforce. There are numerous theoretical models for distributing this wealth: ranging from egalitarian social-welfare systems to labor-leasing arrangements (similar to real estate rentals today), where owners collect revenue generated by the autonomous units they own (whether humanoid robots or cloud-hosted models). In practice, this scenario could drift toward extreme wealth concentration or toward poverty eradication and resource optimization. With honest pessimism, I find it unlikely that AI will be spontaneously adopted in the short term to foster equal wealth distribution. However, I see two scenarios that could force such a shift: The Consumption Collapse: As working-class income plummets, demand for products and services that fund the wealthy class dries up. Economic stagnation would eventually compel large-scale wealth redistribution interventions. The Erosion of Meritocracy: A series of socio-political shifts facilitated by the reality that labor is executed almost entirely by machines, rendering the traditional ethos of meritocracy largely meaningless. Warnings from Detroit: Become Human For a compelling exploration of this tension, look at the narrative of the video game Detroit: Become Human. Released shortly before the widespread adoption of ChatGPT, the game uses biorobotics as a core plot device to explore the cracks in an automated society: Unemployment and Social Decay: We witness the story of a taxi driver falling into debt and substance abuse after his livelihood is eliminated by autonomous vehicle fleets. Automation and Healthcare Dilemmas: A police officer loses his son on the operating table after a car crash because the surgeon on call was incapacitated under the influence; an android attempted the surgery in his place and failed. Objectification and Lack of Prevention: A robot is involved in an accident leading to a human fatality, and authorities treat the unit as disposable hardware rather than investigating root causes to prevent future failures. Even without physical androids walking among us yet, generative AI and decision-making algorithms already exhibit the potential to trigger identical social friction points. Bias, Transparency, and Governance Another critical challenge is algorithmic bias—a problem that exists today but carries far greater stakes in a society where public services lack human alternatives. How do we ensure a public-service AI isn't calibrated by its developers to sway election outcomes, favor specific contract approvals, or alter benefit allocations? If such an AI is developed by a private vendor, model creation and training datasets will demand unprecedented transparency. If built in-house by public agencies, strict oversight is essential to prevent manipulation across changing political administrations or individual misuse. Ultimately, viewing AI merely as a "copilot" for existing jobs misses the broader trajectory: if humans perform purely operational support functions, AI is inherently capable of absorbing those roles. The Glass Half Full On a more optimistic note, maintaining societal infrastructure with a smaller population could mean reduced resource consumption and fewer human labor hours. AI can also accelerate solutions to long-standing global challenges: Medical Breakthroughs: Accelerating vaccine development for rapidly mutating viruses like HIV. Recently, AI models successfully identified novel antibiotic candidates by analyzing the chemical structures of existing medications and calculating viable derivatives—a crucial development given global concerns around antibiotic resistance. Quantum Computing: Accelerating quantum development (which raises an intriguing question: could current AI architectures leverage quantum computing to become exponentially more accurate while consuming significantly less energy?). Conclusion The conclusion to this reflection is as open-ended as its opening: I have no certainty about where we will end up, but one thing is clear—we won't remain static, and we won't reve