Zcash protocol passes AI security audit, but AI raises broader crypto security concerns
Quick Look
- Zcash's protocol has passed a security audit conducted by Anthropic's Claude Mythos AI, finding no serious vulnerabilities.
- However, the increasing use of AI in cybersecurity is also raising concerns about potential exploitation by threat actors in the broader crypto industry.
AI-generated summary
Zcash founder Zooko Wilcox said a security audit by Anthropic's Claude Mythos artificial intelligence model found no serious vulnerabilities in the privacy-preserving cryptocurrency's protocol.
Requested by Shielded Labs, a Swiss-based non-profit supporting the development of Zcash, the AI security audit did not find “any more serious bugs” in the Zcash protocol, according to a Saturday X post by Wilcox.
On June 3, Zcash developers temporarily suspended Orchard transactions after discovering a vulnerability in the shielded pool. Functionality was restored later that day through an emergency upgrade.
The issue stemmed from a four-year-old forgery bug in the Orchard shielded pool that was discovered by security researcher Taylor Hornby with the help of Anthropic's Claude Opus 4.8 model. The Zcash Foundation said there was no evidence that the vulnerability was exploited, nor was there any unauthorized value creation detected, while user privacy was unaffected.
Source: Zooko Wilcox
AI models spark crypto security concerns
While developers are using new AI models to identify vulnerabilities, the technology is simultaneously raising security concerns across the crypto industry.
On Tuesday, Anthropic released the first public version of its Claude Mythos model, Fable 5. The company said last month that the Mythos model uncovered more than 10,000 high or critical-severity vulnerabilities in “systemically important software,” leading to concerns about whether it should be publicly released.
The company said users that Fable 5 was “made safe for general use” and has safeguards that reroute some topics, such as cybersecurity, to a different model, Claude Opus 4.8.
On Friday, Anthropic said it suspended access to its Fable 5 and Mythos 5 AI models due to a US government export control directive citing national security concerns.
Related: Recovery hopes fade as Kelp DAO hacker launders nearly all $220M in stolen funds
The proliferation of these new AI models has shifted the cybersecurity playing field in favor of the threat actors, causing a “vulnerability apocalypse” that is fueling a resurgence in decentralized finance (DeFi) hacks, Mitchell Amador, the CEO of bug bounty platform Immunefi, told Cointelegraph in a recent interview.
Crypto hacks surged to $634 million in April, the highest monthly value since the Bybit hack led to about $1.4 billion in losses in February 2025, according to DefiLlama data.
