Google DeepMind AI Finds New Weaknesses in Legacy RSA-1024 Encryption
In a landmark cybersecurity development, Google DeepMind announced on November 6, 2026, that its AI agents—working under human supervision—had uncovered two significant weaknesses in legacy 1024-bit RSA encryption, a widely deployed cryptographic standard in past decades.
The findings focus on two types of attack:
- Digital signature forgery: DeepMind’s AI agents developed a theoretical method estimated to be approximately 1,600 times more effective than the previous best-known approach for 1024-bit keys.
- Bulk key factoring: The agents discovered an algorithm that reduces the average cost of factoring each additional 1024-bit RSA key by roughly 30 times, while using about 3.5 times less CPU time than earlier methods.
Modern RSA deployments using 2048-bit or larger keys remain secure against these specific findings. The research targets older, shorter key lengths that most organizations have already retired.
DeepMind presented the work as an example of how AI agents can accelerate the scientific process of stress-testing encryption—a task that previously could take human research teams months or years. The findings were published in two working papers dated November 5 and 6, 2026.
Source: Crypto Briefing — Google DeepMind’s AI Agents Expose Two Weaknesses in Legacy 1024-Bit RSA