Humans and AI
AI as cognitive prosthesis, deskilling, algorithmic bias, and the governance choice ahead — how AI reshapes the way people think.
The previous module followed bounded minds into an environment built to capture their attention, and into the crowds that can make them wise or foolish. This one follows them one step further — into a partnership with machines that increasingly do their thinking with them. The agent that decided alone (Module 3), then strategically (Module 4), then under the pressure of the attention economy (Module 5), now decides alongside an artificial one. The question is what that partnership does to the human side of the pair — and who gets to set its terms.
Human-AI Futures: Augmentation, Deskilling, and the Choice Ahead
AI as “System 3” — an external cognitive prosthesis — is already here. AI systems augment human memory (search, retrieval-augmented generation), human attention (notification filtering, summarization), and human reasoning (code assistants, medical diagnosis support). In clinical trials, AI-assisted dementia patients showed 31.7% improvement in task completion. Chess centaurs — human-AI teams — outperformed both humans and AI alone, leading Garry Kasparov to observe that “the process matters more than the components.”
But augmentation carries a shadow: cognitive surrender. When GPS replaced mental maps, spatial reasoning skills measurably declined. When calculators became universal, mental arithmetic atrophied. When AI handles first-draft reasoning, the risk is that human judgment — the capacity to evaluate, challenge, and override — weakens through disuse. The paradox: the better AI performs, the less practice humans get at the skills needed to oversee it.
Algorithmic bias makes the governance question urgent. Stanford researchers found in 2025 that AI skin cancer detection systems were half as accurate for Black patients — not because the algorithms were deliberately biased, but because training datasets underrepresented darker skin tones. Pulse oximeters, approved by regulators, systematically overestimate oxygen levels in patients with darker skin, leading to delayed treatment during COVID-19. The Workday hiring algorithm lawsuit alleged that an AI screening tool discriminated against applicants over 40 and people with disabilities — at a scale no human hiring team could match.
The pattern is consistent: AI amplifies rather than transcends. It inherits bias from training data (which reflects historical discrimination), scales it (one algorithm applied to millions of decisions), and automates it (humans defer to algorithmic authority). The same architecture that enables life-saving diagnosis enables systematic discrimination. The difference is not technology — it is governance.
Thomas Sowell’s distinction between unconstrained and constrained visions offers a framework. The unconstrained vision — embodied in projects like Stargate ($500 billion in AI infrastructure) — assumes that optimization at scale solves coordination problems, that more compute yields better outcomes, and that the gains will diffuse broadly. The constrained vision — embodied in the EU AI Act (enforcement beginning August 2026) — assumes that concentrated power corrupts, that algorithmic decisions in high-stakes domains require transparency and oversight, and that human judgment must remain in the loop for value-laden decisions.
Neither vision fully prevails. The US pursues fragmented regulation (NYC, Illinois, Colorado passing separate AI laws); the EU pursues comprehensive frameworks; China pursues state-directed AI development. The outcome depends not on which technology wins, but on which governance structures societies choose — and for whom they are designed. (For the full governance story, see Politics and Governance.)
AI amplifies rather than transcends human nature. Whether algorithms become tools of liberation or concentration depends on governance choices, not technological capability.
Adjustable Depth
Choose your level of detail on human-AI futures.
AI as cognitive prosthesis (System 3) augments human memory, attention, and reasoning. Chess centaurs — human+AI teams — outperform either alone. But augmentation creates dependency: when AI handles first-draft thinking, the human capacity to evaluate and override weakens through disuse.
Algorithmic bias is not a technical glitch but a structural problem. AI inherits bias from historical training data, scales it to millions of decisions, and automates it so that humans defer to algorithmic authority. The same architecture enables both medical breakthroughs and systematic discrimination.
Two visions compete: the unconstrained vision (build massive AI systems, trust emergent optimization) and the constrained vision (regulate, require transparency, keep humans in the loop). Neither fully prevails globally — the US, EU, and China are pursuing different governance models.
The core insight: AI amplifies human nature rather than transcending it. The outcome depends on governance design — who sets the rules, who is accountable, and whose interests the system serves.
The cognitive surrender phenomenon has historical precedent but new urgency. London taxi drivers who use GPS show measurable hippocampal changes compared to those who navigate mentally (Woollett & Maguire, 2011). Medical residents who rely on diagnostic AI show slower development of clinical intuition. The calibration problem is severe: users tend to either over-trust AI (automation bias — accepting incorrect AI recommendations) or under-trust it (algorithm aversion — rejecting correct AI recommendations), with few achieving the nuanced calibration that centaur models require.
Stanford’s 2025 AI Index documented persistent accuracy gaps across demographic groups in medical imaging, hiring algorithms, and credit scoring. The Workday lawsuit (filed 2023) alleged that AI screening tools produced disparate impact against older workers and disabled applicants — discrimination at a scale impossible for human recruiters. The EU AI Act classifies such systems as “high-risk” and requires conformity assessments, human oversight, and bias auditing before deployment.
The delegation games framework (Riedl & De Cremer, 2025) identifies three independent dimensions of human-AI collaboration: cooperation (willingness to work together), alignment (shared objectives), and calibration (appropriate trust levels). Most failures stem from miscalibration rather than misalignment — users trust AI in domains where it fails and distrust it in domains where it excels.
Generative bubbles (LSE research, 2025) represent a new epistemic threat: AI systems that personalize not just content selection but content generation, creating individualized information environments that are increasingly difficult to verify against shared reality. Unlike filter bubbles, which select from existing content, generative bubbles create content tailored to individual belief systems — a qualitative shift in epistemic risk.
Sowell’s constrained/unconstrained framework, originally from A Conflict of Visions (1987), maps onto AI governance with striking precision. The unconstrained vision (techno-optimism, move-fast-and-break-things, Stargate Project) assumes that human nature is perfectible through technology. The constrained vision (EU AI Act, algorithmic accountability) assumes that concentrated power inevitably corrupts and that institutional checks are essential regardless of technological capability.
These four modules have traced a single thread: from individual bounded rationality, through strategic interaction and institutional design, through the algorithmic reshaping of attention and collective intelligence, to the partnership between human and machine cognition. The pattern is consistent — constraints become structures become power. The cognitive limits that make nudging possible also make manipulation possible. The network structures that enable cooperation also enable cascading failure. The AI systems that augment human capability also concentrate control.
The question is not whether these dynamics exist — they are features of complex systems. The question is whether we design institutions that channel them toward resilience or allow them to drift toward fragility. The remaining modules explore this question in specific domains: history, economics, AI, the digital economy, and governance.