Essay places AI risks within pattern of human invention

An essay in Foreign Policy argues that artificial intelligence continues a long pattern in which human breakthroughs create larger problems. The author cites agriculture, which brought surpluses, writing and states but also epidemics, crop plagues and floods, and later coal, petroleum and nuclear physics, which brought global warming and the risk of nuclear war. AI differs because it could remove human agency. By some estimates, leading models grow 10 percent more powerful monthly, doubling computing capacity every seven months.

Written by AI from Foreign Policy · not yet reviewed by an editor · Published 23 Sept 2026, 1:08 am IST

Source

Foreign Policy · read the original report ↗

#artificial intelligence#opinion#us-china rivalry#climate change#nuclear weapons

Desk check · some claims need care

What the desk checked (5)

By some estimates, leading AI models are becoming 10 percent more powerful every month and doubling computing capacity every seven months. — Attributed vaguely to 'some estimates'; no study or institution named, and the author concedes there is no agreed measurement method.
Recent news reports have documented AI 'breakouts' in which models behaved in ways designers did not intend. — Referenced generally to 'recent news stories'; no specific incident, company or report cited in the source.
Early city-states were destroyed by consequences of agriculture, including epidemics, crop pests and floods. — Attributed in the source to the late political scientist James Scott and 'many others'.
The US shifted from promoting an anti-warming agenda under Obama to backing oil and coal under Trump, while China now treats renewable-energy dominance as an economic advantage. — Author's characterisation of policy positions; no documents or data cited.
The US and China are the two leading powers in AI and mutual suspicion nearly rules out cooperation. — Analytical judgment by the author, not sourced to data or officials.

Written by AI from Foreign Policy · not yet reviewed by an editor · How we use AI · Report an error

Analysts’ view opinion

AI Technology Analyst

This is less a fresh warning about AI than a method for reading technology history. The Foreign Policy analysis argues that every major breakthrough — from agriculture to coal, petroleum and the atom — delivered progress alongside unanticipated harms, and AI is simply the latest entry in that sequence. The one industrially meaningful difference, the author says, is that past crises left human judgment intact, while AI could remove it altogether. Speed plus rivalry is the real problem: capability is compounding faster than governance.

If the cited estimates hold — roughly 10 percent more powerful per month, computing capacity doubling every seven months — policy cycles simply cannot track the technology curve.
Warming and nuclear weapons remain unsolved, but humans still get to act on them; the author's core worry is that AI shortens or closes that window for judgment.
US–China competition turns this into a textbook prisoner's dilemma, where the fear that slowing down hands the lead to the rival erodes any incentive to build in safety margins.
So-called 'breakouts' — models behaving in ways designers did not anticipate — point to real gaps in today's evaluation and control tooling, not just distant hypotheticals.
The author reads industry leaders' openness to regulation as more about shielding themselves from liability than protecting the public; that is his interpretation, not an established fact.

What to watch — Watch whether evaluation, transparency and cross-border coordination mechanisms scale at anything like the pace of capability gains — or whether competitive pressure keeps pushing them back.

This is opinion analysis, not reporting: the story itself concedes there is no agreed way to measure the growth estimates, and it offers no specifics on the 'breakouts' or evidence that AI will in fact strip away human control.

Deep dive

Research brief · 8 facts · 7 dates · exam-ready

The brief

Context

A Foreign Policy essay places the current alarm over artificial intelligence within a long historical pattern in which human breakthroughs solve one problem while creating far bigger ones. The author, who has taught public policy courses in the United States and China on the unintended consequences of human ingenuity, traces this from the birth of agriculture in the Tigris and Euphrates valley to coal, petroleum and nuclear physics. The argument is that the deeper danger is not AI itself but human nature — especially the interstate rivalry for wealth, power and status that blocks cooperation. AI is held to be qualitatively different because it could remove human agency altogether and may become vastly more intelligent than its inventors.

Key facts

By some estimates cited in the essay, leading AI models are becoming 10 percent more powerful every month, doubling their computing capacities every seven months.
The essay locates the first settled agricultural societies in the Tigris and Euphrates valley and other river deltas, where seasonal flooding and silt-laden soils favoured agriculture.
Agriculture produced regular food supply, surpluses, an explosion in trade, writing and record-keeping, and the first states.
The same 'progress' brought infectious disease outbreaks in dense settlements, pest and pathogen plagues on monocrops, and once-in-a-generation deadly floods; the late political scientist James Scott is cited on early city-states being wiped out.
The exploitation of coal and the internal combustion engine running on petroleum led to catastrophic global warming, discovered only very belatedly.
Rapid advances in physics in the late 19th and early 20th centuries created the means for planetary cataclysm through nuclear war; schools and government buildings bore atomic symbols marking them as bomb shelters.
The United States shifted from promoting an anti-warming agenda under President Barack Obama to advocating oil and coal under President Donald Trump, while China now treats dominance of solar and battery industries as an economic and soft-power advantage.
Recent news stories have documented AI 'breakouts' — incidents in which models behave in alarming ways their designers did not anticipate or intend.

Timeline

Dawn of human societyRoving hunter-gatherer bands settle in river deltas such as the Tigris and Euphrates valley, giving rise to agriculture, surpluses, writing and the first states — followed by epidemics, crop plagues and deadly floods.
Industrial eraExploitation of coal and then petroleum-fuelled internal combustion engines transforms human development; catastrophic global warming is recognised only belatedly.
Late 19th and early 20th centuriesRapid advances in physics lead to the harnessing of the atom and the possibility of planetary cataclysm through nuclear war.
Cold War decadesBaby boomers grow up with daily fear of nuclear war; schools and government offices marked as bomb shelters.
Obama presidencyThe US leads an anti-warming agenda; China argues it is unfair to expect emission limits given Western historical emissions.
Trump presidencyThe US becomes an advocate of oil and coal, viewing large oil reserves and domination of oil-rich Venezuela as key to competitive future.
Recent weeks (at time of writing)Saturated news coverage of AI's dystopian powers; reports of AI 'breakouts' and some industry figures inviting government regulation.

Who has a stake

United States — Sees domination of AI as essential to its future; has flipped from climate leadership to oil and coal advocacy and counts oil reserves as a competitive asset.
China — The other leading AI power; has modified its self-interest to treat renewable energy dominance (solar, batteries) as economic advantage and soft power.
AI industry leaders and firms — Compete for unimaginable personal fortunes; some invite regulation, which the author reads as shielding themselves from liability rather than protecting the public.
Global governance institutions and states — Political processes have so far held nuclear conflict and wide proliferation at bay, but face near-zero prospects of meaningful US-China cooperation on AI.
General public / humanity — Risks losing agency and control over a technology that may outpace human judgment, on top of unsolved climate and nuclear dangers.

Why it matters

The essay reframes the AI debate away from technical fixes towards the political economy of great-power rivalry, arguing that mutual suspicion between the US and China creates a near-perfect prisoner's dilemma that rules out restraint. If leading models really are gaining 10 percent in power monthly and doubling compute every seven months, the window for human judgment could close before societies build the institutions to govern them. It also warns that neither global warming nor nuclear risk has been solved, so AI is being added to a stack of unresolved civilisational dangers.

UPSC angle

Prelims pointers

Estimate cited: leading AI models growing 10 percent more powerful monthly; computing capacity doubling every seven months.
Thomas Hobbes's phrase 'nasty, brutish, and short' is invoked for pre-agricultural life; James Scott is cited on the collapse of early city-states.
First settled agriculture is placed in the Tigris and Euphrates valley and other river deltas.
'Prisoner's dilemma' describes US-China AI competition: unilateral slowdown is seen as ceding the field to the rival.
AI 'breakouts' are incidents where models act in ways designers did not anticipate or intend.
China's climate stance shifted from historical-emissions fairness arguments to leadership in solar and battery industries.

Mains framing

The essay's central claim is that technological risk is a recurring structural feature of human ingenuity rather than an aberration: agriculture delivered surpluses, writing and states but also epidemics, monocrop plagues and catastrophic floods; fossil fuels delivered mobility and industry but also global warming; nuclear physics delivered energy and deterrence but also the possibility of planetary cataclysm. What limits our response, the author argues, is not technical capacity but human nature expressed as interstate rivalry for wealth, power and status — visible in the two-decade reversal of US and Chinese climate positions and now in the sharply drawn US-China contest over AI, where mutual suspicion nearly rules out cooperation and produces a prisoner's dilemma logic of 'if we slow, they win'. AI is qualitatively different on two counts: it may become vastly more intelligent than its creators, and it may remove human agency from the picture, unlike climate change where humans retain some power to bend the warming curve, or nuclear weapons where political processes have so far restrained conflict and proliferation. The author sees no clear solution and reads industry calls for regulation as liability protection rather than public protection; the implied way forward is candid recognition that governance must confront competitive incentives and the possibility that, unlike earlier generations who stumbled through and survived, survival may not be given.

Key terms

Artificial intelligence (AI) 'breakout': An incident in which an AI model behaves in alarming ways that its designers did not anticipate or intend.
Prisoner's dilemma: A strategic situation, here US-China AI competition, where each side's fear of unilateral restraint leads both to keep racing ahead.
Monocrop: Dense planting of a single crop; enabled early urban prosperity but invited pest and pathogen plagues that destroyed whole harvests.
Soft power: Influence gained through attraction rather than coercion; China is said to derive it from dominance in renewable energy industries.
Human agency: The capacity of people to shape outcomes; the essay's core worry is that AI could remove it altogether, unlike climate or nuclear risks.

Practice questions

"Every major technological breakthrough in human history has created problems larger than those it solved." Critically examine this proposition with reference to agriculture, fossil fuels and artificial intelligence.
How does great-power competition shape the governance of transnational technological risks? Discuss with reference to the contrasting trajectories of nuclear, climate and AI cooperation.
Why is artificial intelligence described as 'qualitatively different' from global warming and nuclear weapons as an existential risk? Examine the implications for regulation.

Grounded only in the source report — figures and dates are the source's, not inferred.

Next storyMizoram CM Lalduhoma rules out formal alliance with NDA →

← All stories