A new report published by The Quantum Hub in September breaks down the generative AI value chain into eight distinct actor classes and five functional control attributes. The framework aims to identify which actors have the technical ability to prevent, mitigate, or respond to AI-related risks, addressing gaps in India’s current regulatory approach, according to medianama.com.
The report, titled Functional Taxonomy of Actors in the AI Value Chain, categorizes actors such as physical infrastructure providers, cloud access providers, foundation model developers, and end users. It maps these against controls like model layer influence and runtime inference control, producing a matrix that clarifies who can actually manage AI harms. This approach challenges India’s existing IT Act framework, which broadly classifies many different businesses as intermediaries without considering their actual control capabilities.
India’s current AI regulations impose obligations on entities that may lack the ability to comply effectively, resulting in compliance costs without enhancing safety. The report argues that assigning responsibility to actors without matching control undermines regulatory safeguards. By providing a clearer mapping of control and responsibility, the framework could help refine policy to focus on actors who can meaningfully influence AI outputs and risks.
The report was released in September by The Quantum Hub and is available on medianama.com. It offers a foundational tool for policymakers to rethink AI governance by aligning obligations with technical control, potentially influencing future regulatory updates in India’s AI sector.