On 5 November 2025 India’s Ministry of Electronics and Information Technology released the India AI Governance Guidelines. Unlike the EU, which sets binding obligations through dedicated legislation, India has chosen a “principles first” route — establish a governance framework by guidance, then decide whether to legislate as practice matures.
The core orientation
The guidelines are built around four ideas: safety, trust, inclusion and accountability. The specific principles cover transparency, fairness, accountability and risk assessment, and they flag that deployment of high-risk AI systems requires careful evaluation.
Notably, the guidelines also touch on copyright — both the use of copyrighted material in training data and whether generative AI outputs can themselves be protected. There is no international consensus on this, and India’s decision to fold it into a governance framework shows emerging markets are reserving institutional space for the generative AI copyright debate.
How it differs from the EU path
The EU AI Act uses risk tiering as its skeleton, with explicit prohibitions and high-risk obligations — a “hard law” path. India’s guidelines are a “soft law” path, relying on principle-based guidance and industry self-regulation, with weaker direct constraints on companies in the short term.
But soft law is not without effect. Guidance becomes a reference in enforcement and can inform how courts assess whether a company met its duty of care. See our notes on multi-jurisdiction compliance.
Three implications for companies going global
First, India needs its own file. India has no unified AI statute yet, but its data protection law and sectoral rules are taking shape. Entering India cannot rely on copying an EU or Southeast Asian template — a separate rule map is needed. See our multi-jurisdiction service.
Second, assess high-risk scenarios early. The guidelines treat “careful evaluation before high-risk deployment” as a principle, which means companies deploying in healthcare, finance, law enforcement or recruitment should prepare risk assessment documentation.
Third, sort out copyright and training data boundaries. Training data provenance and output ownership will be asked about in India too. Companies should specify authorisation scope and use limits in data procurement contracts — see our cross-border data service.
A longer view
India will host the AI Impact Summit in February 2026, becoming the first Global South country to host a global AI summit. The guidelines and the summit preparation reinforce each other: India wants a voice in how AI governance rules are formed. See our note on the India AI Impact Summit and the China–ASEAN AI governance workshop.
For companies going global, the “principles first” path means more regulatory uncertainty but also more flexibility. Building governance principles into product design early is cheaper than adapting after the rules harden.
This article is based on India’s AI Governance Guidelines and public reporting. It is for general reference only and does not constitute legal advice.