Governance should be viewed as a strategic advantage in the artificial intelligence (AI) era, rather than a constraint on innovation.
So said Khusela Sangoni-Diko, member of Parliament and chairperson of the Portfolio Committee on Communications and Digital Technologies.
Sangoni-Diko delivered the keynote address – titled: “AI governance, public trust and national competitiveness” − at yesterday’s Naspers/Prosus AI Governance and Policy Symposium 2026.
The symposium, which brought together industry players and leaders, legal and government experts, academia and SMMEs, sought to examine AI governance and the evolving policy landscape.
It emphasised the importance of responsible governance in driving economic opportunity and national competitiveness.
The portfolio committee chairperson argued that governance is often presented as a counterweight to innovation, but the real differentiator for countries seeking to lead in AI will be their ability to build confidence among entrepreneurs, investors and citizens.
That confidence, she said, depends on governance frameworks that enable innovation. “Good governance, therefore, is not the enemy of innovation. In fact, it’s a precondition and one of its greatest enablers.”
South Africa’s AI regulatory environment is yet to be finalised and established, with the national AI policy back to the revision stage after it was withdrawn, owing to several of the academic journals cited in the policy document being “completely fictitious” AI-generated citations.
The Department of Communications and Digital Technologies has said it is targeting this financial year (2026/2027) to finalise the country’s AI policy.
According to Sangoni-Diko, governments across the world are searching for the right AI regulatory approach.
She stressed that there is no universal model, and that there shouldn’t be. “Every country must develop a governance framework that reflects its constitutional values, institutional maturity and developmental priorities.
“To date, only 33 out of 178 countries have enacted AI legislation, and most of these are in the European Union. There are other models like South Korea’s AI Basic Act and China’s generative AI measures that offer valuable lessons, but they also show us where there are gaps.”
Sangoni-Diko explained that Parliament’s oversight committee pinpointed the “critical gaps” in some of the existing legislation, describing it as the “reasoning governance gap”.
“We feel that current laws are good at governing inputs like data quality and privacy, and they’re also good at governing outputs like transparency, labels and the like. But they remain largely silent on what happens in between.
“So, we are not effectively governing the reasoning process of AI systems and how it connects the dots, how it gets from data to decision, and these are critical questions that we believe are a foundation for the trust and confidence we want to build.”
Sangoni-Diko further explained that the committee envisions the state and private sector working together as co-architects of South Africa’s AI governance system. “Our roles obviously are distinct, but they are mutually-reinforcing.
“The age when policymakers could retreat behind closed doors and expect it [regulation] to remain relevant for decades has long passed. AI does not evolve in legislative cycles; it evolves continuously. Governance must therefore become equally adaptive.
“In the past, the pace of innovation outstripped the pace of policymaking; for example, the famous set-top box policy saga. The common trend was that regulatory uncertainty delays technological advancement costs. By the time policymakers had understood and developed regulations for one technological breakthrough, another had already emerged to take its place. So, waiting for the perfect regulatory framework is a luxury that we do not have.”
Way forward
While the withdrawn draft national AI policy proposed setting up multiple oversight institutions, Sangoni-Diko said the portfolio committee believes those in place are “sufficient and just need to be better coordinated” to achieve desired results.
“For our part as Parliament, we’re going to fortify working together with government the existing institutional architecture in tandem with ongoing policy evolution. We have a powerful trio of regulators in the form of ICASA, the Competition Commission and the National Consumer Commission that are already being forced to grapple with the challenges associated with AI products.
“Our focus is on aligning the mandates of these different institutions, and that’s why the establishment of the information communication technologies and media regulators forum is so crucial.
“It brings together ICASA, the Film and Publication Board, the Information Regulator and the .za Domain Name Authority to coordinate their efforts. The intention is to get all of them singing from the same hymn sheet.
“It’s a model of innovation with integrity, which will make South Africa competitive. It positions us as a leader in trust-led AI using the resources and expertise we already have.”
Share
