यह पृष्ठ फ़िलहाल केवल अंग्रेज़ी में उपलब्ध है। मूल्यांकन और उसके प्रश्न भी अंग्रेज़ी में हैं।
Responsible AI
Scores are rules, not a black box.
Every score can be traced to your answers and a published formula.
- Reference
- AIRI-RAI-01
- Published
- 26 Jan 2026
- Last updated
- 30 Sept 2026
- Status
- Published
How scores are produced
Scoring is deterministic. The same answers under the same methodology version always give the same score. No machine-learning model decides a score, a maturity band or a ranking. The formula, weights and bands are published on the methodology page, and each result records the version it was scored under.
Where AI is used
An optional assistant can explain a question or summarise a result in plain language. It is currently switched off. If enabled, it will never change an answer or a score, and the text sent to a provider will exclude personal identifiers.
Human oversight
- Submissions can be reviewed by the Index review team, who may ask for clarification.
- Evidence levels (E0 to E3) show how well each answer is supported, so a claim is never confused with proof.
- No decision with legal or financial effect is made about you by automated means.
Limits you should know about
- Most answers are self-reported. Where evidence is not provided, that is shown, not hidden.
- The Index measures readiness at a point in time, not performance or compliance.
- Scores compare organisations only within the same methodology version.
Challenge a result
If you believe a score is wrong, write to send us a message with the assessment reference. We review it and reply within thirty days.