If organizations want to see real business benefits from their AI investments, customers need to trust AI. Systematic public mistrust of AI can only be overcome if we are able to answer questions from customers, regulators, and other appropriate stakeholders about how this technology works. Using blockchain-based accountability provides an achievable and operational path to accountability and enforcement. FICO has developed a private blockchain that automates documentation and standards in model development. This approach has accelerated time to market through AI and analytics innovations, but has also helped us continue producing new models. Blockchain has reduced support issues and model recalls by over 90%. Making this system work was more of a human challenge than a technical challenge. They learned that it is important to start with standards and then develop technology. Making the system user-friendly was non-negotiable. That repeating quick wins is essential. The need to build repositories to keep large AI assets in alternative storage. They knew they needed a capable IT team to handle the maintenance demands of this system.
