Causum Launches AI Governance Platform

- The EU AI Act will be fully applicable on August 2, 2026, requiring comprehensive technical documentation and risk management systems for high-risk AI applications. - Under the CCPA, new rules for automated decision-making technology (ADMT) took effect on January 1, 2027, mandating pre-use notices, risk assessments, and opt-out options for consumers. - The Mars® platform's architecture is based on "ontologically-structured graphs" which allows it to create "enforceable structures" from organizational knowledge, a method designed to bridge the gap between older expert systems and newer, less predictable large language models. - Causum's approach focuses on "decision-time" governance, aiming to approve or reject an AI's decision *before* it is executed, rather than monitoring and flagging decisions after the fact. - The platform utilizes an open-source connector ecosystem to integrate with existing databases and systems in a read-only fashion, allowing it to work with a company's current infrastructure. - For developers, the EU AI Act necessitates a focus on data quality and bias mitigation in training data, especially for applications that could be classified as "high-risk," such as those used in hiring or other significant life decisions. - The CCPA's definition of ADMT is broad, covering any technology that processes personal information to replace or substantially replace human decision-making, which has significant implications for the design and implementation of features that personalize user experiences or make recommendations. - While specific details on the open-source connectors for Mars are not publicly available, the architectural principle is to use them to unify disparate data sources into causal knowledge structures, which then inform the governance layer.