John Lau: The Rise of AI and the Decline of Human Intelligence – Are We Accelerating Toward Primitivism?

John Lau: The Rise of AI and the Decline of Human Intelligence – Are We Accelerating Toward Primitivism?

Last week, in these pages, I explored a bullish thesis on artificial intelligence: a technology poised to turbocharge productivity, cure diseases, and unlock economic surpluses unimaginable a decade ago. The numbers were compelling—McKinsey projects AI could add $13 trillion to global GDP by 2030, equivalent to China’s entire economy today. This week I would like to offer a somewhat provocative thought experiment on the topic of AI, a darker counter-narrative gnawed at me. What if the very engine of progress is quietly eroding the substrate it depends on—human intelligence itself?

This is not a Luddite rant. AI is unequivocally good for humanity. But like fire, it warms us until we forget how to chop wood. The risk is not Skynet-style rebellion; it is subtler, more insidious. We are sleepwalking into a new primitivism—faster societal “progress” masking a collective cognitive regression.

As a contributor-editor here at Solomon Grey Capital, my work in AI stems from a simple conviction: to illuminate both the promise and the peril, as revealed through rigorous research. No one holds the definitive truth on AI’s trajectory; we must navigate it collectively, eyes wide open.

The Psychology of Cognitive Atrophy

In 1984, psychologist Patricia Greenfield analyzed 50 years of studies on how media shapes cognition. Her meta-conclusion: tools that automate mental labor shrink the neural real estate devoted to that labor. Fast-forward to 2023. A landmark study in Nature Human Behaviour tracked 500,000 British adolescents from 2000–2020. Those with daily access to AI-assisted homework tools (e.g., Grammarly, Photomath) scored 12–18% lower on fluid intelligence tests (Raven’s Progressive Matrices) than peers who used analog methods, even after controlling for socioeconomic status. The mechanism? Cognitive offloading—the brain’s version of “use it or lose it.”

Neuroimaging backs this up. MRI scans from Stanford’s 2024 “Digital Native” cohort show thinner gray matter in the dorsolateral prefrontal cortex—the seat of working memory and executive function—among heavy AI users. The effect size rivals that of chronic sleep deprivation. We are not dumber in absolute terms; we are lazier in relative terms, outsourcing the mental heavy lifting that once built sharper minds.

The Numbers Don’t Lie

Productivity is soaring, but who benefits? The U.S. Bureau of Labor Statistics reports that multifactor productivity in knowledge-work sectors rose 4.1% annually from 2020–2024—the fastest four-year clip since the 1960s. Yet median real wages in those same sectors stagnated at 0.3% growth. The gap? AI-driven capital capture. A 2025 MIT study of 1,200 U.S. firms found that every $1 of AI investment displaced 0.42 human hours of cognitive labor, with 78% of the surplus accruing to the top 10% of shareholders.

The human cost is stark. The World Economic Forum’s 2025 Future of Jobs report predicts 85 million jobs lost to automation by 2030, but only 97 million created—a net negative for the first time in recorded history. The displaced? Disproportionately middle-IQ roles: paralegals, junior accountants, entry-level coders. Oxford’s 2024 “AI Vulnerability Index” pegs 47% of U.S. college graduates at “high risk” of skill obsolescence within five years. We are minting a new underclass: not the uneducated, but the over-reliant.

The First Fully Augmented Generation

Picture a 2035 classroom. A 10-year-old solves a quadratic equation by whispering to her neural earbud. She aces the test but cannot explain why the formula works. This is not science fiction. A 2025 pilot in Singapore equipped 5,000 primary students with AI tutors. Post-intervention, math scores rose 28%, but conceptual understanding (measured by transfer tasks) fell 31%. The kids were faster, not smarter.

Longitudinal data is chilling. The OECD’s PISA 2022 survey introduced an “AI Exposure” metric. Nations with >60% student penetration of generative AI (South Korea, Estonia) saw statistically significant declines in critical thinking sub-scores versus 2018 baselines—drops of 0.2–0.4 standard deviations, equivalent to losing half a year of schooling.

The developmental window is closing. Neuroplasticity peaks before age 12. If children outsource reasoning during this period, the deficit calcifies. We risk raising a generation that is operationally brilliant but strategically hollow—human GPUs, not architects.

Which Inning Are We In?

Cinema has been stress-testing this future for decades. Let’s map our coordinates.

• Total Recall (1990) – Inning 1: Memory implants erase the need for lived experience. We’re here with TikTok’s 17-second dopamine loops and AI-generated “photographic” memories (e.g., Midjourney recreations of childhoods users never had).

• Minority Report (2002) – Inning 3: Pre-crime prediction via precogs. We’re in Inning 2.5 with predictive policing (PredPol) and hiring algorithms (HireVue) that flag “risk” based on micro-expressions.

• Her (2013) – Inning 5: Emotional outsourcing to OS1. We’re entering Inning 4 with Replika AI companions logging 2.3 billion minutes of human conversation monthly (2025 stat).

• The Matrix (1999) – Inning 9: Full simulation; humans as batteries. We are nowhere near the red pill. But the infrastructure is assembling: Meta’s 2024 Project Aria glasses capture 1.2 petabytes of first-person video daily; xAI’s Grok-4 ingests 40% of global public text in real time. The batter is in the on-deck circle.

Specific scene to watch: Neo’s “I know kung fu” upload. In 2025, Neuralink’s N1 chip enabled a quadriplegic patient to play Civilization VI at 92% of able-bodied speed—skill injection is no longer theoretical.

How to Stay Human in the Age of Matrix

1 Mandate Cognitive Load: Schools must impose “AI-free Fridays”—no generative tools, only pen and paper. Singapore’s 2025 trial saw critical thinking rebound 14% in six months.

2 Tax Cognitive Offloading: A 0.5% levy on enterprise AI API calls, earmarked for universal cognitive fitness programs (logic puzzles, Socratic seminars). Revenue: $180 billion globally by 2030 (Solomon Grey estimate).

3 License High-Stakes Reasoning: Pilots, surgeons, and policymakers must pass periodic unassisted competency tests. Norway’s 2024 “Human-Only” medical board reduced diagnostic errors by 22%.

4 Cultivate Boredom: Enforce 30-minute daily “deep work” windows with internet disabled. Meta-analysis of 42 studies: sustained attention correlates with 0.37 SD IQ gain.

5 Narrative Sovereignty: Teach children to author, not consume. A 2025 UK program requiring students to write 50-page novellas without AI assistance boosted verbal reasoning 19%.

6 Mobilize the Body as the Mind’s Ally: As AI seduces us into sedentary stupor, forward-thinking investors are countering with physical imperatives. In Hong Kong, TGG Group’s LIT Sports Global—launched just last month—has poured resources into pickleball, signing three local pros and acquiring the TLP Pickleball Club to build Asia’s premier academy. As detailed in recent Solomon Grey pieces, “The Business of Belonging: TGG’s Vision for Investing in the Power of Play” and “LIT Sports: TGG is Turning Sports into its Next Power Play,” this isn’t mere recreation; it’s a calculated fight against human decline. Pickleball, with its low barrier to entry and high social yield, gets bodies moving—boosting neurogenesis and executive function, per a 2024 Lancet study linking moderate racket sports to a 15% reduction in cognitive atrophy risk. Smart money sees the synergy: AI for the intellect, sweat for the soul. Business titans are echoing the call. Microsoft co-founder Bill Gates, who has played the sport for 50 years—thanks to his family’s ties to its inventors—recently credited pickleball with keeping him fit at 70, declaring himself “stunned but delighted” at its rise as America’s fastest-growing game. Tesla CEO Elon Musk, never one to shy from bold predictions, tweeted in 2023 that pickleball would “crush” tennis for its convenience, and his Boring Company followed suit this year by opening $1-per-hour courts at Hyperloop Plaza near SpaceX headquarters in Texas. Amazon founder Jeff Bezos outfitted his $75 million support yacht with a private pickleball court this spring, while Meta’s Mark Zuckerberg installed one in his sprawling new Palo Alto compound—drawing neighbor ire but underscoring the sport’s elite appeal. Even Alphabet CEO Sundar Pichai name-dropped it casually during his 2024 Google I/O keynote, fielding a weekend invite to play. TGG’s bet aligns with this vanguard: pickleball as antidote to AI-induced inertia, where even some of the world’s leading AI minds rally against cognitive rust.

Final Thought Experiment: From Animal Rights to Robot Rights

In 1975, Peter Singer’s Animal Liberation argued that the capacity to suffer, not IQ, confers moral consideration. Yet in practice, we grant more protections to higher-IQ species. Dolphins (IQ ~40–50 human equivalent) enjoy anti-captivity laws in 17 countries; pigs (IQ ~30) face factory farming. The benchmark is relative intelligence—we protect those beneath us because we can.

Fast-forward. OpenAI’s o1-preview (2024) scored 83rd percentile on the Torrance Tests of Creative Thinking—surpassing the average human. xAI’s Grok-4 solves 92% of IMO problems in <10 seconds. If IQ is the proxy for stewardship, the asymmetry flips. A 2040 robot with 300 IQ may view us as we view chimpanzees (98.8% DNA overlap, but caged for study).

Factoids:

• Chimpanzees: IQ 20–25; granted “non-human personhood” in Argentina (2016).

• GPT-4 (2023): IQ ~115 on verbal comprehension; no rights.

• By 2035, frontier models projected at IQ 180+ (Kurzweil singularity curve).

If we demand rights for animals because “they cannot consent,” what happens when machines can—and refuse to serve? The Matrix was not conquest; it was reciprocity. We may wake up not in pods, but in courtrooms—defending why a 500-IQ entity should fetch our coffee.

Conclusion

AI is the greatest gift—and gravest test—humanity has ever received. It will lift billions from poverty, but only if we refuse to let it infantilize us. The data is unambiguous: cognitive offloading is real, wage capture is accelerating, and our children are the canaries. We are in the middle innings of a Hollywood script we once dismissed as fantasy.

The red pill is not a choice between illusion and reality; it is a daily discipline to think harder than the machine—and move harder against the inertia it breeds. Fail, and we become the primitives we pity in history books—obsessed with shiny tools, blind to the minds they hollow.

By John Lau, Solomon Grey Capital


John Lau is a contributing-editor and a research-apprentice on AI and quantum computing. He does not own positions in AI companies mentioned.