Artificial intelligence has managed to solve a fluid mechanics equation that stumped human experts for nine decades, doing so in just 88 hours. What initially appeared to be a triumph for modern science quickly triggered uncomfortable questions about fair play, digital power, and ownership. A university professor accused the world's most prominent artificial intelligence company of quietly copying his team's original research to cross the finish line first. The incident has exposed deeper fears across the globe regarding how technology firms handle human ideas, art, and intellectual property.

A massive digital swarm tackles an old puzzle

Navier-Stokes equations explain how liquids and gases flow, affecting everyday phenomena from weather patterns to ocean currents. For 90 years, mathematicians lacked a full theoretical proof explaining these complex physical movements. To crack the conundrum, OpenAI assembled a massive swarm of roughly 10,000 autonomous digital agents powered by an unreleased model. Over four days starting on 1 September 2026, the artificial intelligence network exchanged nearly three million messages. The system burned through staggering computational power, consuming billions of text output tokens and costing up to 22.5 million US dollars. By 5 September, the digital swarm resolved two key mathematical statements required for a full solution. The achievement targeted a famous prize carrying a one-million-dollar reward from the Clay Mathematics Institute.

Quiet research copied by a tech giant

The celebration was swiftly overshadowed by serious allegations from New York University mathematics professor Tristan Buckmaster. Working alongside Anthropic mathematician Levent Alpöge in a private capacity, Buckmaster had spent months pursuing the same mathematical problem. The pair chose an unusual and rarely attempted mathematical angle involving a smooth force, a path almost no other researchers were following. To build their proof, they relied heavily on OpenAI’s coding tool, known as Codex. Buckmaster revealed that word of their progress reached OpenAI in early September. Shortly after, the firm deployed immense computing resources to attack the exact same path. Buckmaster alleged that an OpenAI researcher suggested removing Alpöge’s name from the credit and warned him against jeopardising his career. He raised alarms that OpenAI models trained on private user interactions with Codex might have absorbed and reproduced their specific reasoning.

OpenAI defends its independent discovery

OpenAI publicly acknowledged the concurrent work of the two mathematicians, describing their findings as remarkable. The company firmly denied accessing any private files or user data to reach its breakthrough. Representatives stated that the project began independently after hearing industry rumors that two major mathematical problems were close to being solved. While acknowledging it could not entirely rule out that de-identified usage data had influenced its underlying models, OpenAI insisted its final mathematical proofs differed significantly from those created by the human duo. The firm added that it had no intention of claiming the one-million-dollar Millennium Prize bounty for its result.

Famous writers and news outlets join the fight

The dispute over mathematical equations reflects a much wider battle over how technology companies collect human knowledge. Major news organisations, including The New York Times, along with celebrated authors like John Grisham and George R.R. Martin, have launched federal lawsuits against OpenAI and Microsoft. Creators argue that tech corporations illegally scrape millions of copyrighted books and news articles without permission or payment. Authors maintain that generative tools dilute literary markets and undermine human creative incentives. In response, technology firms argue in court that analysing statistical language patterns represents transformative learning and does no financial harm to original writers. American judges remain divided, with some calling the technology transformative and others warning that unchecked artificial intelligence could flood markets with competing content.

From patent disputes to global technology battles

Traditional legal frameworks across the globe have long required human intellect to grant patents and copyrights. Patent offices worldwide recently rejected applications for inventions created by an artificial intelligence system called DABUS, ruling that only human beings qualify as legal inventors. Meanwhile, international tensions over digital property are escalating rapidly. United States security agencies issued formal advisories accusing Chinese technology firms of committing industrial-scale theft by using a technique known as distillation to extract secret capabilities from American models. Cybersecurity experts also report that international hackers are actively targeting valuable artificial intelligence source codes and server quotas. In developing nations such as India, where artificial intelligence is rapidly entering schools, legal systems, and hospitals, existing laws offer limited guidance on machine-made creations. As machines increasingly generate complex ideas, global institutions are struggling to protect human creators while encouraging technological progress.