The UAE Cyber Security Council and Open Innovation AI announced on 17 September a cybersecurity AI model for organisations that need to keep sensitive security data inside private systems. The partners have not said when companies can use it, what it will cost or how its performance has been independently tested.
The announcement names vulnerability assessment, secure-code analysis and remediation, security-patch analysis, incident investigation, response support and security-advisory review as intended tasks. It also says the model can generate structured outputs for security operations centres, enterprise systems and related cybersecurity platforms.
The private-deployment claim is central to the launch. A security team investigating a breach may work with source code, incident logs, network details and internal advisories that it would not want copied to a public AI service. The partners say the model can operate in controlled and sovereign environments, allowing an organisation to retain control of the data and workloads it uses.
That claim does not establish that any deployment is secure. Security still depends on configuration, access controls, telemetry and operational practice. The announcement also does not name the model, publish its architecture or model size, identify the proprietary training data, provide a download route, list customers or set out licensing and pricing terms.
In May, the Cyber Security Council, Open Innovation AI and Cisco announced a National AI Test and Validation Lab for assessing AI systems for security, safety and trustworthiness. The September launch does not say whether the new cybersecurity model has been assessed there. The lab and the model should therefore be treated as related UAE initiatives, not as evidence that the model has passed an independent evaluation.
For organisations in Dubai and elsewhere in the UAE, the immediate position is clear: this is an announced technology, not a product they can currently sign up for. The next information that would make it actionable is a technical paper, evaluation results, access terms, pricing or named deployments. Until then, the launch establishes the model’s intended role, not that it is ready to replace specialist security judgment.