India · National guidance
NITI Aayog Responsible AI principles, put into practice
NITI Aayog's Responsible AI for All papers (2021) set out seven principles for AI in India and how to put them into operation. They're voluntary, but sector regulators and public buyers increasingly expect teams to show how each principle is met.
What the NITI Aayog RAI asks of AI systems
The obligations most relevant to bias, risk, and accountability, summarised from the NITI Aayog Responsible AI principles.
- P1Safety and reliability
- Systems behave as intended and risks of harm are managed.
- P2Equality
- People in similar circumstances are treated equally.
- P3Inclusivity and non-discrimination
- Systems do not unfairly exclude or discriminate against groups.
- P4Privacy and security
- Personal data is protected throughout the lifecycle.
- P5Transparency
- Design and functioning can be explained and audited.
- P6Accountability
- Clear ownership for decisions the system makes or supports.
- P7Positive human values
- Systems reinforce rather than undermine human values.
Evidence FairMind helps you produce
FairMind links technical evidence to NITI Aayog RAI obligations for a human reviewer to accept. It is not legal advice and does not decide compliance.
- Bias and non-discrimination evaluations across groups relevant to Indian deployments
- Named owners and reviewer decisions for every finding
- Transparency records: model version, intended use, and known limitations
- A principle-by-principle evidence index with gaps left visible
NITI Aayog RAI questions
Are the NITI Aayog Responsible AI principles mandatory?
No. They are national guidance, not law. Sector regulators such as the RBI draw on similar principles, so documenting against them prepares you for binding rules.
How do I show inclusivity and non-discrimination?
Run bias evaluations on the groups your system affects, set thresholds before the run, fix what fails, and retest. FairMind links each of those steps into one reviewable record.