I built a tool to prove a human reviewed an AI decision (caneni.net)

A new tool named Caneni has been developed to provide verifiable proof that a human reviewed and made decisions based on AI output. The platform aims to fill a critical gap in AI governance by documenting human oversight with structured, timestamped records that detail who reviewed the AI's decision, what factors were considered, and the decision made. This is particularly pressing given recent legal precedents, such as Israel's Supreme Court ruling that criticized a municipality for blindly relying on AI-generated decisions for a child with special needs, highlighting the need for accountability in algorithmic decision-making. The significance of Caneni lies in its potential to bolster compliance with emerging regulations, including Israel's upcoming Privacy Protection Law Amendment and the EU's AI Act, both emphasizing human oversight in high-risk AI systems. This tool not only supports regulatory requirements but also shifts liability and accountability to organizations utilizing AI outputs, as underscored by Google's notification regarding service activity. By creating a reliable method for recording human decisions, Caneni positions itself as a pioneer in a new category of AI oversight, fostering a culture of accountability in a rapidly evolving technological landscape.