Aztec Foundation CEO Zac Williamson said advanced AI could expose flaws in Tierra, underscoring a security challenge for Ethereum’s privacy network as only selected protocol components have undergone formal verification.
Williamson made the comments in an interview recorded Nov. 9. He said he “can’t rule out” bugs being found by future AI models. Tierra is also known as Alpha V6.
Formal verification uses mathematical reasoning to check that code matches its design. Aztec has applied the process to components it considers most critical to security, while Williamson called checks across the full protocol a long-term goal.
Williamson said Aztec’s in-house system runs audits continuously. He said V5 reviews assisted by AI exposed critical vulnerabilities in cryptographic libraries, prompting the team to formally verify those subsystems.
A separate flaw in V5’s proving system could have let a forged proof pass verification. The flaw’s exploitation status is unverified. Tierra includes a fix for it.
During a major Aztec upgrade, users must take assets out of the old version before depositing into its successor, Williamson said. An Ethereum withdrawal can reveal a concealed double-spend in the older version, he said, adding that the flaw “can’t extend to the new version of Aztec.”
Williamson said Aztec operates without a security council. He identified full-protocol verification as the project’s long-term objective.

![[Satoshi Index] Top Traders’ 24-Hour Long Share Jumps From 8% to 80%](https://wfuxxvxhdynkhcrlavdu.supabase.co/storage/v1/object/public/article-media/articles/bf3e85c85a204d3062bfa959be95c7a1670447520fa477293f612119f9f198b2.webp)

