Data governance is the foundation for responsible AI at scale, and getting the fundamentals right determines whether institutions capture AI value or inherit AI risk.
So says Dhesen Ramsamy, group chief technology and data officer at Old Mutual, who is scheduled to present at the ITWeb AI Summit 2026 on 22 April at The Forum in Bryanston.
Ramsamy will outline how sovereignty and governance enable, rather than constrain, innovation. He will elaborate on why data quality determines AI performance and why data governance determines AI trustworthiness.
Speaking to ITWeb ahead of the event, Ramsamy said: “Without proper data management (quality standards, lineage tracking, bias detection), AI initiatives fail regardless of model sophistication. South Africa's National Data and Cloud Policy (June 2024) emphasises data sovereignty and localisation. In practice, many organisations discover their AI readiness depends on first addressing data governance for regulatory compliance. The regulatory imperative often drives the necessary foundation.”
In a national context, he said, data governance balances three imperatives: sovereignty (data subject to local laws), economic value capture and innovation enablement.
“The framework should support legitimate use while protecting citizen rights and national interests,” Ramsamy added. “Managing and protecting data in a national context means recognising that data is part of a country’s critical infrastructure – as essential as energy, transport or financial systems. It requires us to treat citizen and organisational data as a national asset that must be governed responsibly, secured consistently and used ethically.”
Part of his presentation will address the challenges of managing and protecting data in this national context, including infrastructure and skills shortages as key barriers to digital transformation.
Ramsamy said: “Some reports indicate that only a small percentage of African AI talent currently has access to the computational power required for innovation. While South Africa has a significant number of data centres compared to other African nations, the continent's overall share of global AI-capable compute capacity is relatively small.”
He also noted that while the national AI policy framework was released in October 2024, businesses continue to adopt the technology without a ratified AI Act in place.
“Business adoption outpaces regulation. Some institutions may deploy autonomous agents for high-stakes decisions (trading, customer service) without sector-specific accountability standards. Financial services require algorithmic explainability, audit trails and human oversight standards. The current vacuum creates risk as deployment accelerates. We need sector-specific frameworks with practical enforcement mechanisms,” Ramsamy continued.
He said regulation alone remains insufficient, and the real challenge is building capability – moving from AI consumers to AI builders through technology transfer frameworks, compute infrastructure investment and skills development at scale.
For more information and to register, visit ITWeb AI Summit 2026.
Share