Creator here. With the SEC and EU mandating real-time, immutable audit trails for AI after the recent Nasdaq crash, I saw a critical need for a simple, open-source standard that anyone can implement. This is a proof-of-concept for a hash-chained log. The idea is to create a deterministic "proof-of-process" for any AI decision. The next step is to integrate zk-STARKs so companies can prove their AI's integrity without revealing the proprietary models or sensitive data. The GitHub repo is linked in the article. Happy to answer any questions and looking for feedback/contributors.