Images are used for representative purposes. |Photo courtesy: Getty Images/iStockphoto
Recognizing that Artificial Intelligence (AI) will impact a wide range of sectors, the Union Government Committee mandated to develop AI guidelines will help enforce AI regulations and ensure effective governance in emerging countries. It recommended that an inter-ministerial committee be established. AI ecosystem.
The advisory group of the Union Ministry of Electronics and Information Technology’s IndiaAI Mission, led by Chief Scientific Adviser, has sought public feedback on the report released by the AI Guidelines Sub-Committee on Monday (January 7, 2025). The report suggests a “whole-of-government, coordinated approach to enforcing compliance and ensuring effective governance as India’s AI ecosystem evolves.”
The panel’s report is notable because it is the most detailed look yet at the government’s approach to AI in India. The center appears to be seeking to maximize returns from its AI investments while having a light regulatory framework that acts as guardrails against potential harm.
Multidisciplinary expertise
The commission should establish a “technical secretariat” in the IT ministry, separate from the interministerial AI coordinating committee and governance group that oversees AI in India, and invite delegates from other ministries and regulators. He recommended that. This, along with other research, will “integrate multidisciplinary expertise” and help map actors within the AI ecosystem and assess risks, the committee recommended.
This report outlines the following principles for AI governance: Transparency of AI systems with “critical information about development” and capabilities. Accountability by developers and adopters of AI systems. Design AI systems to be safe, reliable, and robust. Privacy and security of AI systems. Fairness and non-discrimination. Human-centered values and “do no harm”. Inclusive and sustainable innovation to “equitably share the benefits of innovation.” “Digital by design” governance to “leverage digital technology” to operationalize these principles.
life cycle approach
To operationalize these principles, policymakers should use a life cycle approach, the report says. This means that AI systems need to be considered at various stages of development, deployment, and dissemination, where distinct risks can arise. Second, we need an “ecosystem view” of AI actors. A “holistic” view of AI in practice requires consideration of all data principals (those whose data will be used), data providers, AI developers, deployers, and end users. .
A “command and control” governance strategy “may not adequately facilitate the growth and expansion of this sector,” the committee said. Instead, this report suggests a technology-enabled digital governance system. “This could allow players to manage their in-chain debt, allowing for lightweight and incrementally scalable regulatory controls.” On issues such as deepfakes and cybersecurity, the report reiterates past government positions on this issue, that existing laws provide the legal framework to address these issues.
“Enabling self-regulation”
The report warns of the need for “transparency and accountability across the AI ecosystem” and that “categorizing systems purely based on computational power or data parameters may not be effective.” It warns that there may be sector-specific risks no matter what AI you look at. systems such as healthcare and banking. The report recommends a “whole-of-government” approach, citing the “potential inefficiencies and gaps of a piecemeal approach” (highlighted in the report).
The commission, the report recommends, should “apply and strengthen existing legislation to minimize the risk of harm from the use of AI,” harmonize interagency efforts, and improve the AI sector. It says there is a need to “enable empirical self-regulation” between AI sectors, as well as identifying gaps between AI sectors. Create a policy environment that fosters “support for multiple regulations,” “useful use cases,” and fosters the deployment of “responsible AI applications” in individual sectors.
issued – January 8, 2025 12:30am IST
