World War IV and the Law: Data, Dependence and the Quiet Erosion of Human Judgment

Sep 26

7 min read

A prophecy worth revisiting

There is a line almost every law student has heard at some debate or moot: Einstein supposedly said he did not know what weapons World War III would be fought with, but World War IV would be fought with sticks and stones. The most commonly cited source is an interview by Alfred Werner in Liberal Judaism (April–May 1949), recorded in the Einstein Archive and reproduced in Alice Calaprice's The New Quotable Einstein (2005). Even so, fact-checkers caution that the line may not be an exact record of his words, and the earliest accounts appear to be anecdotal.

Einstein was imagining a civilisation bombed back to the Stone Age. I want to propose a quieter version. What if the next great conflict does not destroy our cities, but hollows out our capacity to think, and leaves us in a world where we still have every tool imaginable but have lost the judgment to use them? The foundations of that war are being laid right now, and the legal profession is not a bystander.

Front one: data as the new strategic reserve

Wars have always been fought over resources: land, oil, sea lanes. The resource of this century is behavioural data. A handful of technology companies, mostly headquartered in a few developed economies, now hold more information about how people think, shop, vote and feel than any intelligence agency of the Cold War could have imagined.

Indian constitutional law saw this coming. In Justice K.S. Puttaswamy (Retd.) v. Union of India, (2017) 10 SCC 1, a nine-judge bench recognised privacy as a fundamental right under Article 21. Yet it took years for Parliament to give that right working machinery. The Digital Personal Data Protection Act was passed in 2023, and MeitY released the final DPDP Rules through Gazette Notification G.S.R. 846(E) dated 13 November 2025, ten months after the draft rules came out in January 2025. Even now, implementation is phased over eighteen months, running to May 2027.

That lag matters. A law that arrives after the data has already been collected, modelled and monetised is protecting a field after the harvest. Whoever holds the richest datasets trains the most capable AI systems, and whoever controls those systems increasingly shapes the information that citizens, legislators and judges see. That is the strategic high ground of World War IV.

Front two: outsourcing the mind

The second front is internal. We are handing over not just our data, but our thinking.

The most cited recent evidence comes from the MIT Media Lab. In Your Brain on ChatGPT, researchers divided participants into three groups (LLM users, search engine users, and a "brain-only" group with no tools), with 54 people completing the first three sessions and 18 completing a fourth. Using EEG, they found that brain-only participants showed the strongest and most widely distributed neural networks, search engine users showed moderate engagement, and LLM users showed the weakest connectivity. The authors call the result "cognitive debt."

A second study looked at professionals rather than students. Researchers from Microsoft and Carnegie Mellon surveyed 319 knowledge workers, who shared 936 first-hand examples of using generative AI at work. The key finding: higher confidence in GenAI was associated with less critical thinking, while higher self-confidence was associated with more critical thinking. The researchers also observed that AI use shifts work from gathering information to verifying it, from solving problems to integrating AI responses, and from doing tasks to supervising them.

A fair reading requires honesty about limits. The MIT team itself notes that the paper had not been peer-reviewed as of June 2025, so its conclusions should be treated with caution and as preliminary, and the sample was small and geographically narrow. It would be an overstatement to say that nobody uses their mind anymore. But the direction of the evidence is hard to ignore: when we trust the machine more than ourselves, we think less.

Front three: the attention economy and the youth

If AI is outsourcing our reasoning, short-form video is fragmenting our attention. In 2024 Oxford University Press chose "brain rot" as its Word of the Year, defining it as the supposed deterioration of a person's mental or intellectual state, especially as a result of overconsuming trivial or unchallenging material, now particularly online content. Oxford noted that the term gained traction among Gen Z and Gen Alpha, who amplified it through the very social media channels said to cause it.

The science is catching up with the slang. A 2025 meta-analysis in the American Psychological Association's Psychological Bulletin pooled data from 98,299 participants across 71 studies and found that heavier short-form video use was associated with poorer cognition, with attention and inhibitory control showing the strongest links. Notably, these findings held across youth and adult samples and across different platforms, which includes Instagram Reels and YouTube Shorts.

Again, a caveat: these are correlations, not proof of causation. Critics have pointed out that the cognitive analysis relied on only 14 studies, all conducted in China. But for a country like India, with one of the youngest populations on earth, even a correlation of this size deserves the attention of policymakers. A generation trained to swipe every fifteen seconds is not a generation trained to read a 200-page judgment.

Front four: when the courtroom joins the war

This is where the issue stops being a lifestyle concern and becomes a rule-of-law concern. Courts are where societies go to think slowly and carefully. What happens when the court itself outsources its thinking?

India's first open experiment came in 2023, when in Jaswinder Singh @ Jassi v. State of Punjab Justice Anoop Chitkara of the Punjab and Haryana High Court asked ChatGPT about the jurisprudence on bail where assailants had acted with cruelty. The judge was careful to note that the reference was not an opinion on the merits and was meant only to present a broader picture of bail jurisprudence. It was a transparent, cautious use. What followed was less so.

In the Bengaluru bench of the ITAT, Buckeye Trust v. PCIT had ruled for the tax department in a Rs 669-crore dispute using three fake Supreme Court citations and one fake Madras High Court citation; the department's representative had reportedly used ChatGPT, and the bench copied the citations without verification. In the Delhi High Court, Greenopolis Welfare Association v. Narender Singh was withdrawn after opposing counsel showed its citations were fabricated, including paragraphs 73 and 74 of Raj Narain v. Indira Nehru Gandhi, a judgment that runs to only 27 paragraphs. In KMG Wires Pvt Ltd v. National Faceless Assessment Centre, the Bombay High Court found that an assessing officer had cited three non-existent decisions to justify a Rs 22.66 crore addition to a company's income.

The judiciary has started to push back. On 19 July 2025, the Kerala High Court issued its Policy Regarding Use of Artificial Intelligence Tools in District Judiciary, drawing a firm line that AI may not be used "to arrive at any findings, reliefs, order or judgment." The policy applies to district judges, their staff, and any interns or law clerks working with them.

The Supreme Court has now gone further. In February 2026, taking up an Andhra Pradesh trial court order built on non-existent precedents, it declared that a decision based on fake AI-generated judgments is not an error but misconduct with legal consequences. Then on 2 July 2026, in the Essel Infraprojects insolvency matter, a bench of Justices PS Narasimha and Alok Aradhe set aside NCLT and NCLAT orders after finding that three of the six judgments relied on did not exist and three carried invented paragraphs or a wrong case title. The Court compared hallucinated case law to "the release of methyl isocyanate in the province of law and justice", a deliberate echo of Bhopal that should unsettle every practitioner. The Court's draft AI rules, released in June 2026, would bar AI from deciding cases and require lawyers to disclose when they use it.

Why this leads to a faulty future

Put the four fronts together. Data is concentrated in a few hands. Citizens are losing the habit of sustained attention. Professionals are trading thinking for supervising. And the institutions meant to be the last line of careful reasoning are themselves vulnerable to confident, fluent, fabricated output.

Policy is only as good as the minds that make it. A legislature that cannot read a bill, a bar that cannot verify a citation, and an electorate that cannot hold a thought for longer than a Reel are not ingredients for a stable democracy. No bombs are required. The weapons of World War IV are convenience and speed.

What lawyers can do

I am not arguing for abandoning AI. Used well, it can clear backlogs, translate judgments and widen access to justice. But a few principles seem non-negotiable to me:

Verify every authority against a primary source. The Supreme Court has made it clear that "the AI did it" is no longer a defence.
Disclose AI use in pleadings and orders, even before the draft rules become final.
Treat AI as a first draft, never a final answer. The Microsoft–CMU findings suggest self-confidence in your own expertise is what keeps critical thinking switched on.
Protect deep work. For law students especially, reading full judgments rather than summaries is not old-fashioned; it is cognitive self-defence.

Einstein worried we would be reduced to sticks and stones. The more realistic risk is that we keep all our technology and lose the judgment to govern it. That war is winnable, but only if we notice it has begun.

References

A. Werner, "Einstein at 70," Liberal Judaism (Apr–May 1949), cited in A. Calaprice, The New Quotable Einstein (Princeton UP, 2005), p. 173; see also Snopes, "Albert Einstein said 'World War IV would be fought with sticks and stones'" (2018).
Justice K.S. Puttaswamy (Retd.) v. Union of India, (2017) 10 SCC 1.
Digital Personal Data Protection Act, 2023 (No. 22 of 2023); Digital Personal Data Protection Rules, 2025, G.S.R. 846(E), 13 Nov 2025; PIB Backgrounder, "DPDP Rules, 2025 Notified" (17 Nov 2025).
N. Kosmyna et al., "Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task," arXiv:2506.08872 (2025), https://arxiv.org/abs/2506.08872
H-P. Lee et al., "The Impact of Generative AI on Critical Thinking," Proceedings of CHI '25 (ACM, 2025), doi:10.1145/3706598.3713778
L. Nguyen et al., "Feeds, feelings, and focus," Psychological Bulletin 151(9): 1125–1146 (2025), doi:10.1037/bul0000498
Oxford University Press, "'Brain rot' named Oxford Word of the Year 2024" (2 Dec 2024).
Jaswinder Singh @ Jassi v. State of Punjab (P&H HC, 27 Mar 2023, Chitkara J.), reported in Verdictum (29 Mar 2023).
High Court of Kerala, Policy Regarding Use of Artificial Intelligence Tools in District Judiciary, HCKL/7490/2025-DI-3-HC (19 Jul 2025).
Buckeye Trust v. PCIT (ITAT Bengaluru, Dec 2024).
Greenopolis Welfare Association v. Narender Singh (Del HC, Sept 2025).
KMG Wires Pvt Ltd v. National Faceless Assessment Centre (Bom HC, Oct 2025).
Gummadi Usha Rani v. Sure Mallikarjuna Rao, CRP No. 2487 of 2025 (AP HC, Jan 2026); Supreme Court order dated 27 Feb 2026.
Supreme Court of India, judgment dated 2 Jul 2026 (Essel Infraprojects insolvency / J&K Bank), summarised in A. Bansal, "10 cases that show Indian courts have an AI hallucination problem," MediaNama (3 Jul 2026).