GPT-6 Astra Solves a WWI German Radio Cipher

AI can now solve complex historical ciphers that were previously impenetrable, potentially unlocking vast archives of lost information.

The AI model GPT-6 Astra has successfully decoded a World War I German radio cipher that had remained unsolved for over a century. The cipher, a German radio message transmitted on November 27, 1918, was encoded using the ADFGVX method. Researchers had previously decoded many ADFGVX messages, but over a dozen eluded efforts. GPT-6 Astra identified "TRUPPENVERSCHIEBUNG" as the encryption keyword, a key hypothesized to have been used by the Germans starting in December 1918, which may explain why previous attempts failed for this earlier message. The decoded message states, "AN ENGLISH CRUISER ARRIVED AT SEVASTOPOL ON THE ?4TH AN ALLIED SQUADRON FOLLOWS ON THE 26TH." Verification using historical logs confirmed the arrival of the English cruiser HMS Canterbury in Sevastopol on November 24, 1918, and an allied squadron on November 26, 1918, aligning with the decoded message.

This achievement by GPT-6 Astra demonstrates a significant leap in AI's capability for complex pattern recognition and contextual inference, extending beyond current computational linguistics tasks. The model's ability to hypothesize and test keyword variations, even those not in common use at the specific historical transmission date, signifies a more sophisticated understanding of cryptographic challenges. This implies future AI models could tackle other long-standing historical mysteries, from lost texts to uncrackable codes, by leveraging deep learning to perform analysis previously reserved for human experts with decades of specialized knowledge.

The successful decoding of the ADFGVX cipher has implications for both historical research and the cybersecurity industry. For historians, it opens avenues to potentially deciphering other historical documents and communications previously considered beyond reach, offering new insights into past events. For cybersecurity, it highlights the increasing power of AI to break complex encryption, underscoring the need for continuous evolution in cryptographic methods to stay ahead of advanced analytical tools.