HashSmash Contest Pits AI Agents Against Core Cryptographic Hash Functions
Eigen Labs and Shielded Labs launch HashSmash, an open competition testing whether AI can crack widely used hash functions.
Two prominent blockchain organizations have joined forces to stress-test the cryptographic foundations of the modern internet. Eigen Labs and Shielded Labs — the development entity behind the Zcash privacy protocol — have launched HashSmash, an open multiplayer competition designed to measure how effectively artificial intelligence can advance attacks against widely used cryptographic hash functions.
The contest, hosted on the Yukon platform, invites participants from anywhere in the world to deploy AI agents against four hash functions: SHA-256, SHA-3, BLAKE3, and Poseidon. These algorithms sit beneath much of the encryption infrastructure that secures online communications, financial transactions, and blockchain networks. By opening the challenge to the public, organizers aim to surface vulnerabilities — or confirm resilience — under adversarial AI pressure.
Read more FTC Settles With Martial Arts Franchisor for $1.85M Over Deceptive Claims →
Transparency is central to the initiative. Every result generated during the competition will be published and subjected to review by professional cryptographers, creating a crowdsourced audit of AI-driven attack methods. That open-review structure distinguishes HashSmash from private internal security assessments and gives the broader research community visibility into emerging threats.
The timing reflects growing concern within the security industry about AI's potential to accelerate cryptanalysis. As large language models and autonomous AI agents become more capable, cryptographers and protocol designers are increasingly asking whether encryption standards built for the pre-AI era will remain adequate. HashSmash frames itself as a proactive attempt to answer that question before adversaries do.
Continue reading at Cryptocurrency
