Has the Flynn Effect Peaked? Intelligence in an Age of AI and Global Realignment
The Flynn effect—that remarkable, steady rise in IQ scores observed across the 20th century—may have run its course. After decades of gains averaging about three points per decade, evidence from Scandinavia, the UK, and other developed nations suggests scores have plateaued or even reversed. The question isn’t whether this is happening, but what it means for education, economic competitiveness, and the future of human potential.
James Flynn himself, who documented the phenomenon bearing his name before his death in 2020, was always careful to note he was measuring something more nuanced than raw intelligence. The gains reflected improved abstract reasoning, better nutrition, smaller families, more years in school, and greater familiarity with the cognitive demands of modern life. The tests, in essence, were capturing adaptation to modernity.
If that’s what drove the rise, the plateau makes sense. Nutrition in wealthy countries reached adequate levels decades ago. Family sizes stabilised. Educational attainment, while still improving at the margins, has largely saturated. The low-hanging fruit has been picked.
The data are striking. Norwegian conscript scores rose from 99.5 for the 1962 birth cohort to 102.3 for those born in 1975, then declined to 99.4 for the 1989 cohort. Danish military testing shows similar patterns. In Britain, Flynn’s own 2009 study found that 14-year-olds’ average IQ dropped by more than two points between 1980 and 2008. A 2018 study in the Proceedings of the National Academy of Sciences confirmed that both the rise and the reversal are environmentally caused—not genetic drift, not changing demographics, but something in the conditions of modern life.
The Great Divergence
But here lies a crucial asymmetry that Western commentators often overlook. While the Flynn effect may be reversing in Scandinavia and stagnating in the Anglosphere, it continues apace in much of the developing world. Meta-analyses show that BRIC nations—Brazil, Russia, India, and China—are currently experiencing the greatest gains, averaging 2.9 points per decade compared to roughly two points in wealthy countries and near-zero or negative in some. Studies have documented continuing Flynn effects in Kenya, South Africa, and across Southeast Asia.
This divergence carries profound implications for global competitiveness. The cognitive capital of nations isn’t static. If developed economies are experiencing diminishing returns while emerging markets continue accumulating intellectual capacity, we should expect accelerating shifts in innovation, entrepreneurship, and economic dynamism. The talent pipelines feeding Silicon Valley and the City of London increasingly originate in Hyderabad, Shenzhen, and Lagos. The Flynn plateau in the West may prove less significant than the Flynn continuation elsewhere.
For the Gulf states, where I work, this presents both challenge and opportunity. The UAE, Saudi Arabia, and Qatar have invested heavily in education infrastructure and attract global talent. But sustainable knowledge economies require homegrown cognitive capacity, not merely imported expertise. Whether regional educational investments translate into genuine Flynn gains—rather than credential inflation—remains an open question.
AI: Prosthesis or Crutch?
More troubling explanations for the Western plateau deserve consideration. The elimination of lead from petrol contributed an estimated four to five IQ points to American scores between the 1970s and 2000s—a one-time gain now fully realised. Educational reforms in some countries have shifted emphasis away from abstract reasoning toward practical skills. And then there is the question of how we spend our cognitive time.
The relationship between screen-based technology and cognitive development remains genuinely contested. Some studies suggest excessive screen time displaces cognitively demanding activities—reading, complex play, sustained problem-solving—that once exercised young minds. Research links heavy early childhood screen exposure to slower language development and reduced executive function. Yet a major Oxford Internet Institute study of nearly 12,000 American children found no evidence that screen time impacted brain function or wellbeing. Another study found that time spent gaming was actually associated with modest IQ gains. The honest answer is that we don’t yet know whether digital technology is helping or harming cognitive development—or whether the effects vary dramatically by type, content, and context of use.
But the elephant in the room is artificial intelligence. As we increasingly outsource cognitive tasks to machines—calculation, memory retrieval, navigation, even writing and analysis—are we enhancing human capacity or allowing it to atrophy? The question echoes ancient anxieties. Socrates worried that writing would weaken memory. Critics of calculators feared innumeracy. Perhaps each technological prosthesis costs us something while gaining us more.
Yet AI represents a qualitative leap. Previous tools augmented specific cognitive functions. AI systems increasingly replicate general reasoning itself. When students use large language models to draft essays, solve problems, and synthesise information, they may be learning to direct intelligence rather than develop it. The Flynn reversal might accelerate precisely as AI becomes more capable—not because we’re becoming less intelligent in some absolute sense, but because the cognitive muscles that IQ tests measure are getting less exercise.
There’s an alternative framing. If we measure intelligence by what humans plus their tools can accomplish, the question becomes moot. The relevant unit of analysis isn’t the individual brain but the human-AI system. By that measure, cognitive capacity is exploding. But this raises uncomfortable questions about dependency, resilience, and what happens when the tools fail or access proves unequal.
The Measurement Problem
Perhaps the deeper issue is that IQ tests themselves have become artefacts of a bygone era. They were designed for industrial-age sorting—identifying who could be trained for clerical work, who for manual labour, who for the professions. The skills that matter in a networked, globalised, AI-augmented economy don’t map neatly onto pattern recognition and working memory.
Consider what contemporary challenges actually demand: navigating radical ambiguity, synthesising across disciplines, collaborating across cultures, distinguishing signal from noise in an information-saturated environment, maintaining focus amid engineered distraction. These capacities don’t register on standard IQ metrics. A brilliant content creator, a skilled negotiator, an entrepreneur who senses market shifts before they materialise—none would necessarily score higher than their grandparents on a Raven’s Progressive Matrices test.
My own academic work sits at the intersection of finance and geopolitics—applying option pricing frameworks to alliance behaviour, analysing infrastructure corridors through both economic and strategic lenses. This kind of boundary-crossing synthesis represents a cognitive mode that didn’t exist when IQ tests were designed and that they’re poorly equipped to assess. Perhaps we’re not getting dumber; we’re getting different in ways our instruments can’t detect.
Recent research supports this interpretation. A 2025 Norwegian study found that declines were concentrated in numerical reasoning and word comprehension—domains affected by educational reforms that de-emphasised mental arithmetic once calculators became ubiquitous—while other cognitive dimensions showed different patterns. We may be witnessing not a general decline but a reallocation of cognitive resources.
From Diagnosis to Prescription
If the Flynn plateau represents genuine cognitive stagnation rather than measurement failure, policy responses become urgent. Several interventions merit serious consideration.
First, we should treat cognitive development as a public health issue, not merely an educational one. This means addressing environmental factors—from air quality to endocrine-disrupting chemicals—that may be undermining neurological development. The dramatic cognitive gains from eliminating lead exposure demonstrate the power of environmental intervention. The precautionary principle should apply: we needn’t wait for definitive proof of harm before limiting children’s exposure to substances with plausible cognitive effects.
Second, while the research on screen time remains contested, the precautionary case for protecting children’s attention has merit. The business models of social media platforms depend on capturing and fragmenting human attention. Even if the cognitive effects prove modest, there are opportunity costs when hours that might be spent reading, building, or engaging in complex play are instead devoted to passive consumption. Age-appropriate limits and emphasis on interactive over passive screen use seem prudent pending clearer evidence.
Third, educational systems need recalibration. If AI will handle routine cognitive tasks, education should emphasise what humans do distinctively well—creativity, ethical reasoning, interpersonal connection, physical skill, aesthetic judgment. The premium on rote learning and standardised testing may be exactly backwards. We should be cultivating the capacities machines can’t replicate, not those they’re about to master.
Fourth, and perhaps most importantly, we need better instruments for measuring what we actually care about. IQ tests served their historical purpose. Continuing to rely on them risks optimising for yesterday’s cognitive demands while ignoring tomorrow’s.
The Stakes
The Flynn effect was never inevitable. It was the dividend of specific historical conditions—public health advances, educational expansion, the cognitive demands of industrialisation and urbanisation. If those conditions have changed, so too will the trajectory of human cognitive performance.
But framing this as simple decline misses the complexity. Different populations are on different trajectories. Different cognitive capacities are evolving differently. And the human-machine interface is transforming what intelligence means and how it manifests.
What’s certain is that we can’t afford complacency. Cognitive capacity—however measured—remains foundational to economic productivity, democratic functioning, and human flourishing. If the developed world is experiencing diminishing returns while emerging economies continue gaining, and if AI is simultaneously augmenting and potentially atrophying human cognition, we’re navigating a transition of profound consequence.
The Flynn effect’s peak, if that’s what we’re witnessing, isn’t an ending. It’s an inflection point—demanding that we think more carefully about what intelligence is, how it develops, and what kind of cognitive future we want to build.
