The Artificial Superintelligence Alliance and developer platform Matterhorn have announced a partnership to mitigate security risks from AI-generated blockchain code. Their initiative focuses on Matterhorn‘s “vibe coding” feature, where AI generates full smart contracts from plain language descriptions. The collaboration combines automated analysis, third-party security audits, and human review to vet code before deployment on the ASI:Chain network, aiming to onboard 20,000 developers in 2026 without guaranteeing security.
The Artificial Superintelligence Alliance and developer platform Matterhorn are building tools to reduce risks from AI-generated blockchain code. This addresses vulnerabilities introduced by Matterhorn‘s “vibe coding,” which instantly generates smart contracts from descriptive text.
The platform combines automated analysis, human review, and testing tools to audit smart contracts before deployment. Developers on Matterhorn can connect with third-party security auditors through the platform prior to launching contracts.
“We partner with security audit companies who can offer audit services through Matterhorn for builders on Matterhorn,” said Matterhorn founder Abhinav Ramesh. “We have AI agents as well that do agentic audits, but we absolutely don’t recommend doing just that for mainnet applications.”
The partnership runs on ASI:Chain and aims to onboard 20,000 developers in 2026. The initiative seeks to provide a single environment for building, auditing, and deploying decentralized applications.
Matterhorn integrates with the ASI:Chain network, developed by the ASI Alliance which includes Fetch.ai, SingularityNET, and CUDOS. The companies are developing “blessed templates” to help build safer contracts for formal verification-based languages.
While the platform accelerates development, it does not guarantee security. “There are absolutely no guarantees of any kind from the Matterhorn team on safety or security,” Ramesh stated.
Khellar Crawford, chief innovation officer of SingularityNET, contrasted the typical “patch-and-pray” approach with their methodology. He explained ASI:Chain uses a “correct-by-construction” architecture based on Rho calculus.
“We don’t guess if an application is safe, we mathematically prove it using spatial behavioral types,” Crawford said. “Before a single line of code ever touches the live network, the math itself guarantees that there will be no deadlocks, no race-condition exploits, and no leaked funds.”
