Some providers selling sovereign AI offer no evidence of it. Omdia research cited by HGS found that 28 per cent of providers disclose no proof of sovereignty, and that some expect enterprises to pay for audits to verify what they are being sold.
HGS, a digital customer experience and business process transformation company, has published a white paper titled AI Sovereignty: What enterprise leaders must get right before the next audit. It sets out a framework for defining sovereignty across five dimensions (data, models, infrastructure, operations, and talent and governance), identifying the controls different AI workloads need, and demonstrating them to an auditor.
The paper argues for a workload-based approach, with controls matched to the applicable legal frameworks, regulatory obligations and the sensitivity of each deployment. Control across the five dimensions varies, with some capabilities kept in-house and others shared with or delegated to third parties.
Rajamani Saravanan, SVP and Head of Engineering at HGS, said: "Most businesses struggle to build AI solutions that can withstand regulatory scrutiny because they have not clearly defined what sovereignty means for their deployments or the controls needed to meet their regulatory obligations."
He pointed to dependence on public cloud infrastructure and proprietary frontier-model APIs, which enterprises may have limited ability to control or verify independently, as a limit on the sovereignty and transparency providers can offer. The white paper can be downloaded from the HGS website.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
