[India] MeitY has released India’s AI Governance Guidelines, a four-part framework shaping how AI will be developed, deployed, and governed.
Here’s a breakdown of its structure.
2/The framework spans 4 interconnected parts:
1️⃣ Seven Sutras the ethical spine of India’s AI governance model.
2️⃣ Six Pillars mapping enablement, regulation & oversight.
3️⃣ Action plan with timelines + institutional roles.
4️⃣ Operational Guidelines aligning industry & regulators for consistent adoption.
7/ The guidelines are drafting with the intent to encourage innovation and adoption while protecting individuals and society from risk ( such as malicious use e.g. misrepresentation through deepfakes, algorithmic discrimination, lack of transparency, systemic risks and threats to national security).
8/ No need for new law to address AI separately.
Existing laws (for e.g. on information technology, data protection***, consumer protection and statutory civil and criminal codes, etc.), can be used to govern AI applications.
It emphasises the need for timely and consistent enforcement of applicable laws to build trust and mitigate harm
9/ Amendment to Copyright Law is required.
To enable the large-scale training of AI models, while ensuring adequate protections for copyright holders and data principals.
10/ Suggests development and adoption of voluntary framework by AI keyplayer to mitigate risks.
Further, it seeks transparency about the AI value chain (which comprise of actors like developers, deployers and users) as it will help regulators design proportionate & effective governance mechanisms.
11/ Techno-legal approaches promoting compliance by design (including for content authentication via watermarking etc. + content traceability) can be applied to support specific policy objectives.
12/ Recognising the definition of AI incidents as given by OECD, the committee suggests creating a national database of ‘AI incidents’, which gives policymakers insights into the real-world risks and harms posed by AI systems.
13/ Adopt a graded liability system based on the function performed, level of risk, and whether due diligence was observed. Applicable laws should be enforced, while guidelines can assist organisations in meeting their obligations.
14/ The recently established AI Safety Institute (AISI) should act as the main body responsible for guiding the safe and trusted development and use of AI in India.
