Corporates eye greater AI control to protect data and trade secrets

This story has significance for readers across Kenya and beyond.

For the first wave of artificial intelligence (AI) adoption in the past four years, businesses had a relatively simple formula: subscribe to OpenAI’s ChatGPT, Microsoft's Copilot or Google's Gemini, upload large volumes of data, automate repetitive tasks and move on.

But as AI becomes embedded in core business operations such as market research, customer relationship management and logistics, some firms are increasingly questioning who controls the intelligence behind the technology.

Increasingly, these companies are deciding that relying entirely on external AI providers may not be enough. Instead, they are investing in what technology executives call 'sovereign AI'. This is building or deploying AI systems in a way that allows an organisation to retain control over its own data, infrastructure and AI models instead of depending entirely on third-party platforms.

The shift is driven by growing concerns about data privacy, intellectual property (IP) and regulation as businesses entrust AI with sensitive internal information.

Sovereign AI allows firms to keep sensitive corporate and customer data within secure private infrastructure, unlike consumer AI tools, which process much of their information in public cloud environments. This helps the companies comply with local privacy laws and data residency requirements.

It also gives companies greater control over how AI models are deployed and reduces dependence on a single technology vendor, while protecting proprietary information from being used to train public or competing AI systems.

According to a recent report by the global IT services and consulting provider NTT Data, the biggest reasons organisations are investing in sovereign AI are to gain a competitive advantage, increase bargaining power and build in-house expertise.

The survey found that 35 percent of chief information officers cited those factors as the main motivation, ahead of meeting country-specific regulatory requirements (26 percent), and simplifying IT environments and reducing operational risk (23 percent).

The trend is mainly prominent in sectors where trust and data sensitivity are critical. NTT Data said public sector organisations lead global interest in sovereign AI, followed closely by healthcare, natural resources—including mining, oil and gas—and manufacturing.

In government, AI increasingly interacts with citizen records, public services and, in some cases, national security systems. In healthcare, it is being used in diagnostics, treatment planning and clinical research, making questions about where patient data is stored and processed increasingly important.

"It all centres around confidence; that the AI and the data that I have is falling in line with the laws of the country and the cultures of the country; that it is being used responsibly in everything that we do," Alan Turley-Jones, Chief Executive Officer of NTT DATA Middle East and Africa, told the Business Daily in an interview.

"Citizens and organisations really want to make sure that the data being used within these AI models is being used responsibly."

Analysts say the concern becomes even greater for companies whose competitive advantage lies in their own information. This is compared to the rest of AI users who might use public AI tools every day without much thought about where their prompts are processed.

For firms in financial services, e-commerce and retail, and digital media and entertainment, exclusive data and advanced information systems drive performance, pricing, and customer loyalty.

"Many corporations' data is their intellectual property. It differentiates them from their competitors," says Mr Turley-Jones. "You've got to make certain that if you are using AI in your corporate environment, that data is being protected.”

“Employees also need to understand what information belongs in protected environments versus what can be used in the public domain."

The conversation is also becoming increasingly relevant in Africa as governments develop AI regulations and businesses accelerate investments in emerging technologies.

Around the world, approaches to sovereign AI differ widely. In the European Union, for instance, investment is often driven by regulatory requirements around privacy and data protection. In parts of the Middle East, national strategy and technological independence are stronger drivers. For African businesses, however, cost is also a major factor.

Building a frontier AI model from scratch requires enormous computing power, thousands of specialised graphics processors, vast amounts of electricity and highly paid AI researchers, making it prohibitively expensive for most organisations.

Instead, many companies are building applications on top of existing open-source models such as Meta's Llama or Mistral, or integrating powerful AI models through application programming interfaces (APIs) provided by firms such as OpenAI and Anthropic.

The strategy allows organisations to retain significant control over their data without shouldering the enormous cost of developing foundational AI systems from the ground up. For Kenya and the wider continent, Mr Turley-Jones argues that the success of sovereign AI will ultimately depend on investments in digital infrastructure.

"AI requires a lot of data and a lot of bandwidth, so we need continued investment in connectivity; data centres are required for these large language models and other AI platforms to run,” he says.

"We can't underestimate the importance of cyber. Ensuring organisations have the right cybersecurity posture is critical to making sure everything we do around AI is done responsibly."

Reporting originally appeared via Business Daily. Read the full source for additional context.