Anthropic AI Confirms Origin of NIST P-256 Curve Seed
Researchers reconstructed the generation of the cryptographic standard to settle long-standing concerns over potential backdoors.
Updated on Oct. 16, 2026 in Artificial Intelligence
Live Poll
Do you trust the security of internet standards established by government agencies in the 1990s?
Anthropic researchers have identified the cryptographic origin of the NIST P-256 elliptic curve seed, settling decades of speculation regarding potential backdoors. Using the Claude AI model, the team reconstructed the generation method in under 30 minutes.
Why it matters
The analysis resolves long-standing concerns regarding the integrity of industry-standard cryptographic curves by confirming they were generated via a specific pseudorandom number generator and Unix timestamp, rather than malicious tampering.
The NIST P-256 curve seed (C49D360886E704936A6678E1139D26B7819F7E90) originated from candidate #1842 of a lrand48/drand48 pseudorandom number generator initialized with the Unix timestamp 871696354.
The players
Anthropic
An AI safety and research company known for developing the Claude series of large language models.
NIST
The National Institute of Standards and Technology, a federal agency that establishes cryptographic standards for U.S. government use.
Jerry Solinas
The researcher who submitted the P-256 and other elliptic curve seeds to the ANSI X9.62 committee in 1997.
Internet Security Research Group
A non-profit organization focused on improving internet security through open-source software and research.
The details
Researchers used the Claude AI model to analyze how 1990s-era PRNGs (pseudorandom number generators—algorithms that produce a sequence of numbers approximating true randomness) behaved during the seed generation process. By reverse-engineering the inputs, the team matched the specific timestamps used to generate various prime and binary elliptic curves. Jerry Solinas originally submitted these parameters to the ANSI X9.62 committee in 1997, leading to their standardization in FIPS 186-2.
Timeline
February 1984: Binary curve seed timestamps.
August 16, 1997, 01:52:34 UTC: Random number generator run started.
Fall 1997: Seeds submitted to ANSI X9.62 committee.
May 1997 - April 1998: Prime curve seed timestamps.
October 16, 2026: Research findings announced.
The Tech Race
This finding settles uncertainty surrounding the cryptographic origins of curves defined in FIPS 186-2. It establishes a new benchmark for using AI models to conduct forensic analysis of legacy cryptographic standards.
Users of internet security protocols relying on NIST P-256, such as those used in TLS/SSL encryption, benefit from increased confidence in the fundamental integrity of these curves. No action is required for developers, as the underlying math of the standard remains secure and unchanged.
The takeaway
The successful use of AI to solve a decades-old cryptographic mystery demonstrates the potential for machine learning in historical data forensics. Security researchers should monitor upcoming audits of other FIPS-standardized curves to confirm they follow the same established generation patterns.
Further reading
Explore more developments in Artificial Intelligence.
Live Poll
Do you trust the security of internet standards established by government agencies in the 1990s?
