AI standards are becoming more than technical paperwork: they can shape what evidence buyers demand and who gets to sell AI-enabled services across borders. Dr Srinivas Padmanabhuni argues that countries in the Global South need the capacity to help write, test and certify those standards, not simply adopt them.
Watch Desk analysis
What happened
In a Global Policy Journal article, Padmanabhuni points to ISO/IEC 42001, a certifiable standard for AI management systems, and argues that standards and audits are helping define what trustworthy AI looks like in practice. India has adopted the standard nationally and has taken part in international standards work, he writes. But adoption alone does not guarantee influence, or the ability to produce credible evidence about real systems.
His prescription has three parts: invest in shared testing facilities, datasets and benchmarks; fund experts from developing countries to take part in standards committees; and pursue mutual recognition so companies are not made to pay for equivalent audits twice. Padmanabhuni is CTO and co-founder of an AI testing and assurance company, and says his own experience informs the argument.
Why it matters
A certificate can show that an organisation has a process for managing AI risk. As Padmanabhuni notes, it does not by itself prove that a particular system is fair or safe. The practical question is who can afford to produce the technical evidence buyers want, and whose auditors and results are trusted.
That matters to exporters, especially smaller firms. If standards become market-entry conditions while testing capacity and certification remain concentrated elsewhere, countries may comply with rules they had little hand in shaping, and pay again to prove they have done so.
Our read
The strongest part of Padmanabhuni’s argument is its focus on the evidence behind a certificate. Standards committees matter, but so do the local-language datasets, testing facilities and independent expertise that make assessments meaningful. A tidy compliance document is not a substitute for knowing how a system behaves.
His recommendations are an argument from an interested practitioner, not a neutral scorecard of the Global South’s readiness. Still, the distinction between adopting standards and having the capacity to shape and test them is useful. The less glamorous infrastructure may be where the real competition is decided.
What to watch
- Whether proposed AI testing facilities receive sustained funding and clear governance.
- Whether developing-country experts gain meaningful roles in international standards committees.
- Whether trade agreements recognise audits across borders, reducing duplicate costs.
Discussion spark: Should countries prioritise building their own AI testing and certification capacity, even if that costs more upfront, or is adopting international standards enough to compete?
Sources and evidence
- AI standards as statecraft: a playbook for the Global South | Global Policy Journal (30 October 2026, 00:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.