THE POLICY EDGE
Reports/Data Releases

8 October 2026

ITU Offers an Eight-Part Test for Putting AI Governance into Practice

The International Telecommunication Union’s new guide helps governments assess whether they have the institutions, infrastructure, skills and safeguards to oversee AI. It identifies India’s mix of digital legislation and sectoral oversight as a case where coordination across regulators and governments matters

Reports/Data Releases image

Key Details

The ITU’s Collaborative Artificial Intelligence (AI) Frameworks and Preparedness Diagnostic is a guide for assessing readiness, not a new regulation or a ranking of countries.

What the Guide Provides

How It Works

Eight areas of preparedness

Examines strategy, institutional coordination, digital infrastructure, data governance, skills, innovation and procurement, societal safeguards, and enforcement.

Evidence-based assessment

Combines document review, interviews and institutional self-assessment. Seven indicators in each area are scored for whether they are absent, partly in place or operating in practice.

A route from assessment to action

Uses the results to identify priority gaps, assign responsibilities and sequence reforms. The guide suggests repeating the diagnostic every two to three years.

An India discussion

Describes India’s approach as led by broader digital legislation and sectoral initiatives, including data protection and financial-sector oversight. It does not score India’s readiness.


The Test Is Whether AI Rules Can Be Used

An AI strategy says little on its own about whether a public agency can safely buy an AI system or a regulator can investigate one. The ITU guide turns that problem into practical questions: Who has authority? Can agencies obtain technical evidence? Do they have skilled staff, reliable data and ways for people to challenge harmful decisions?

Its eight-area diagnostic is designed for national or subnational governments and individual sectors. Scores provide a starting point, but the guide places greater value on finding specific gaps and dependencies than on producing one overall readiness number. Strong computing capacity, for example, cannot compensate for an inability to supervise how AI is used in public services.

Regulators Need to Work Across Their Mandates

An AI system used in lending may raise questions about financial conduct, personal data, cybersecurity and consumer protection at once. The guide therefore calls for clear institutional roles and regular cooperation among the authorities concerned, supported by shared guidance and assessment tools.

It gives telecommunications and digital regulators a substantive role because connectivity, cloud services, computing capacity and reliable power affect both AI use and regulators’ ability to examine it. The guide also treats public procurement as a governance tool: contracts for government AI systems can provide for documentation, audit access, security and accountability.

India’s Question Is How Existing Arrangements Fit Together

The guide describes India as developing AI governance through digital legislation and sector-specific initiatives, rather than a single comprehensive AI law. It points to the Digital Personal Data Protection Act, national AI initiatives and oversight by financial regulators. It also notes that AI use in areas such as health and education raises Centre–State coordination questions.

This is an account of India’s policy landscape, not an ITU assessment that India has passed or failed the diagnostic. The guide’s contribution is a way to examine whether those arrangements can work together when an AI application crosses institutional boundaries.


Policy Relevance

For India, the diagnostic offers a practical distinction between having policies and being able to act on them. A sector-by-sector assessment could show where an existing regulator can oversee AI using its current powers, where it needs technical expertise or access to system records, and where responsibility is unclear. Applying the same test to public procurement and high-impact services would make safeguards part of deployment decisions, not just broad policy statements.

The useful output would be a sequenced set of responsibilities and capacity needs, rather than another aggregate readiness score. The ITU guide proposes a method for producing that evidence; it does not prescribe a particular new Indian regulator or legal model.


Follow the Full Guide Here: ITU, Collaborative Artificial Intelligence (AI) Frameworks and Preparedness Diagnostic

Rethinking Public Policy Through Insight | Inquiry | Impact

Opinion • Grassroots Voices • Policymakers Perspectives • Expert Analysis • Policy Briefs