Gulf banks want AI, but can they keep control of customer data?

Gulf banks are keen to embrace artificial intelligence, but many are still trying to answer a basic question: how can they use the technology without putting sensitive customer information at risk?The issue is becoming increasingly important as banks across the region explore new AI tools that can speed up routine tasks, analyse documents and improve productivity.For Najla Ibrahim Al-Mutawa, Executive Vice President of Strategy and Business Development at QNB, the debate goes beyond efficiency."For banks, the question is not simply whether generative AI can improve efficiency or customer experience, but whether it can be deployed in a way that protects trust, safeguards data, and meets regulatory expectations."That challenge has also created opportunities for companies trying to make AI easier for banks to use safely.Sami Mian, CEO of Blade Labs, says many banks are comfortable with the AI systems themselves, but remain concerned about what information those systems can access."The AI tool may be approved. The cloud may be approved. But the bank still needs to control what the AI is allowed to see."Blade Labs has developed a platform called ZeroH Disclosure, which aims to automatically limit the information shared with AI systems while keeping a record of what data was disclosed and why.For many banks, the biggest challenge may not be the technology itself, but how to control it.Banks need to know that customer names, account details and other sensitive information are protected before data is shared with AI tools.AI and digital transformation adviser Alina Timofeeva says generative AI is forcing financial institutions to rethink how they manage data."In banking, trust is the product," she said. "The question moves from where is the data stored to who can access it, how it is used and who is accountable if something goes wrong."That is particularly relevant in the Gulf, where regulators are pushing ahead with digital transformation while also strengthening rules around data protection, cybersecurity and AI governance.Al-Mutawa said banks are becoming more selective about how they use AI. Low-risk experimentation is being treated differently from applications involving customer data and other sensitive information."Customer data, confidential internal information, financial crime controls, risk models and proprietary business information require much stronger safeguards," she said.Mian says the answer lies in giving institutions more control over what information can be shared with AI systems in the first place.Rather than relying on staff to manually remove sensitive details from documents, the company says those controls can be built directly into the process, allowing only authorised information to be disclosed while creating an audit trail of what was shared.The same thinking is also being applied in Islamic finance, where product approvals often involve multiple stakeholders, including legal teams, compliance departments, auditors and Shariah scholars.Blade Labs is also developing Ask Ali, an AI assistant focused on Islamic finance. The platform is designed to help professionals research standards, review documents and navigate Shariah-related questions while maintaining human oversight of the process.All three say trust is likely to be the deciding factor in how quickly banks adopt AI."The institutions that solve this first will be able to use AI more freely," Mian said. "The ones that cannot prove control will remain stuck in pilots, restrictions and internal approvals."