Artificial intelligence identifies origin of decades old encryption numbers

Cybersecurity researchers at Anthropic used their AI assistant to solve a decades-old mystery surrounding the foundations of modern web security. For years, experts debated the origin of mysterious numbers supplied by the US National Security Agency for widespread encryption standards. By analyzing how software was built in the 1990s, the AI system discovered that the numbers were not built from secret codes, but were generated by a basic computer clock timestamp.

This outcome provides a major sanity check for everyday internet safety. Countless daily digital activities, including secure banking and private messaging, depend on these mathematical formulas to shield user data from hackers. Because official agencies never explained how the original numbers were chosen, critics worried that intelligence services had intentionally planted secret vulnerabilities to spy on web traffic. Proving that the values came from a standard random process confirms that no hidden backdoors were added.

While this discovery resolves a major technical question, it leaves open intriguing administrative questions about why public agencies kept the generation method secret for so many decades. As Anthropic reported, early creators could not remember the exact details, allowing speculation to swirl for a generation. It raises the broader question of whether automated AI systems will now be routinely deployed to recheck other ancient mathematical standards that quietly keep the internet running.