The race to build a quantum-proof digital future just hit a major roadblock. A promising algorithm, HAWK, designed to withstand attacks from future quantum computers, has been pulled from the running after a security model from Anthropic, called Mythos, found a flaw that rendered it broken.
What Mythos found: The flaw that broke HAWK
HAWK was a digital signature scheme—a cryptographic tool used to verify the authenticity of digital messages and documents. It had survived two rounds of rigorous testing by the National Institute of Standards and Technology (NIST), which is evaluating post-quantum cryptographic (PQC) algorithms for official US standards. But during the third round, designed to catch precisely these kinds of vulnerabilities, Mythos uncovered a critical weakness. The flaw was severe enough that the algorithm is now considered broken, meaning it can no longer be trusted for secure communications.
Why this matters for your digital security
Quantum computers, once they become powerful enough, could break many of the encryption methods that currently protect everything from online banking to private messages. Algorithms like HAWK were being developed as replacements. Its withdrawal means one less candidate in the pipeline, potentially delaying the transition to quantum-safe systems. For everyday users, this underscores the urgency of the global effort to secure data before quantum computers arrive.
How HAWK got here: A timeline of testing
HAWK entered NIST’s PQC evaluation process as a candidate for a digital signature standard. It passed the first two rounds, which involved public scrutiny and analysis by cryptographers worldwide. The third round was intended to be the final filter before standardization. Mythos, an AI-powered security model developed by Anthropic, was used to probe the algorithm for hidden weaknesses. The discovery of the fatal flaw in this late stage is a testament to both the difficulty of designing quantum-resistant algorithms and the power of AI in cybersecurity.
Who is affected by this setback
The immediate impact is on the cryptographic community and organizations planning for quantum-safe transitions. Companies and governments that had been monitoring HAWK’s progress now need to look at other candidates. For the broader public, the news is a reminder that the path to quantum security is not straightforward. Every broken algorithm is a lesson that strengthens the remaining ones, but it also highlights the immense challenge ahead.
Anthropic and the developer respond
Anthropic announced the Mythos findings on Monday. The following day, the developer of HAWK confirmed the withdrawal from the NIST process. “The flaw was fundamental,” the developer stated, acknowledging that the algorithm could no longer be considered secure. NIST has not yet commented on the withdrawal, but the agency is expected to continue evaluating other candidates in the PQC program.
What the Mythos attack reveals about AI in security
The Mythos model is designed to find vulnerabilities in cryptographic systems by simulating attacks. Its success against HAWK demonstrates that AI can now play a decisive role in breaking—and by extension, strengthening—cryptographic algorithms. This is a double-edged sword: the same AI tools that can expose flaws can also be used by adversaries to find them first. The incident marks a new chapter in the cat-and-mouse game between security researchers and potential attackers.
Confirmed facts vs what remains unclear
Confirmed: Mythos found a critical flaw in HAWK. The algorithm has been withdrawn from NIST’s third round. The developer acknowledged the break. Unclear: The exact nature of the flaw has not been publicly detailed. Whether other algorithms in the NIST pipeline are similarly vulnerable is unknown. The timeline for NIST’s final selection of quantum-resistant standards remains uncertain.
Risks and balanced view
The withdrawal of HAWK is a setback, but it is also a success for the testing process. The system worked: a vulnerability was found before the algorithm could be standardized. Critics may argue that the reliance on AI for such testing introduces new risks, including the possibility of false positives or missed flaws. Supporters see it as a necessary evolution in cryptography. The broader risk is that the delay in finalizing standards leaves systems vulnerable for longer.
Wider trend: The AI-cryptography arms race
This event is part of a larger pattern: AI is increasingly being used to both attack and defend cryptographic systems. From breaking classical encryption to testing quantum-resistant algorithms, AI is reshaping the field. The Mythos attack on HAWK is a clear signal that the future of cybersecurity will be defined by AI-driven discovery and counter-discovery.
What this means for researchers and organizations
For cryptographers, the lesson is to incorporate AI-based testing earlier in the design process. For organizations planning quantum-safe transitions, the advice is to diversify: do not rely on a single algorithm. Monitor NIST’s ongoing evaluations and prepare for multiple standards. For the general public, stay informed but not alarmed—the process is designed to catch flaws before they become crises.
Future outlook: What happens next
NIST will continue evaluating other PQC candidates, including those in the third round and beyond. The agency may also call for new submissions. The Mythos model will likely be used to test other algorithms. The incident may accelerate efforts to develop AI-assisted verification tools for cryptography. The ultimate goal—a set of quantum-resistant standards—remains on track, but the path is now more complex.
Our Take
The Mythos attack on HAWK is a sobering reminder that quantum-resistant cryptography is still a work in progress. While the withdrawal is a blow, it is also a victory for the rigorous testing process that NIST has championed. The use of AI to break an algorithm that had survived two rounds of human scrutiny is a milestone. It shows that the tools we build to protect the future must themselves be tested by the most advanced systems we have. This is not a failure of the process—it is proof that the process is working.
Frequently Asked Questions
What is the HAWK algorithm?
HAWK is a digital signature scheme designed to be secure against attacks from quantum computers. It was being evaluated by NIST as a potential US standard for post-quantum cryptography.
What is the Mythos security model?
Mythos is an AI-powered security model developed by Anthropic. It is designed to find vulnerabilities in cryptographic systems by simulating attacks.
Why was HAWK withdrawn from NIST testing?
Mythos found a critical flaw that rendered HAWK broken. The developer withdrew the algorithm from the third round of NIST’s post-quantum cryptography evaluation.
Does this mean quantum computers are a threat now?
No. Large-scale quantum computers that can break current encryption do not exist yet. The withdrawal of HAWK is part of the preparation process to ensure future security.
