Loading story…
Trisha Ray, associate director at the Atlantic Council's GeoTech Center, emphasized the growing importance of AI sovereignty as nations seek to control their own artificial intelligence infrastructure, data, and talent. The push for sovereign AI is driven by concerns over national power, trust, and resilience, as countries aim to reduce dependence on foreign technology providers and assert greater control over their digital futures. Ray highlighted that achieving true AI sovereignty is challenging due to the lack of diverse, high-quality datasets and the limited digitization of many languages, which hampers the development of inclusive and effective AI models.
The discussion also explored the broader geopolitical landscape, outlining four possible futures for national AI ecosystems and the foundational elements required for AI sovereignty, such as data, computing power, and skilled personnel. The interview underscored that AI sovereignty is no longer a theoretical concept but a pressing issue at the intersection of geopolitics, technology, and national security, with significant implications for cyberwarfare, fraud management, and the global balance of power.
Track how attackers are adapting to this technology.
2 events from the most recent confirmed update back to the earliest known activity.
Ray said sovereign AI depends on three core components: data, compute, and talent. She also highlighted major barriers, including limited access to diverse, high-quality datasets and the insufficient digitization of many languages, which can hinder inclusive national AI ecosystems.
In an Information Security Media Group interview, Atlantic Council GeoTech Center resident fellow Trisha Ray described AI sovereignty as countries being able to build and control AI using their own infrastructure, data, and talent. She framed it as a practical geopolitical and national security concern tied to trust, resilience, and independence from dominant technology firms and foreign states.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
A Lawfare analysis says the push for **AI sovereignty** is intensifying a long-running dispute over whether unauthorized cyber intrusions into another state’s networks violate sovereignty under international law. The article contrasts a **pure sovereignty** position associated with France and the African Union, an **effects-based** approach reflected in the Tallinn Manual and many Western states, and an **operational-latitude** posture linked to the U.K. and, in a different form, the U.S. It argues that governments have historically preserved room for intelligence collection and offensive cyber operations by resisting a broad sovereignty rule, even as AI systems become more central to national security. The piece warns that this legal posture leaves AI assets such as training data, compute infrastructure, and model weights exposed to low-level but strategically significant intrusions including data theft and data poisoning. Drawing on examples such as the **OPM breach**, **Volt Typhoon** activity, and security cooperation under **AUKUS Pillar II**, it says current thresholds focused on physical damage, loss of functionality, or coercive interference do not clearly cover covert compromise of AI systems. Background references to France’s formal cyber-law position and a U.K. government speech on international law underscore the divide among states as customary international law appears to be moving toward stronger sovereignty protections in cyberspace.
Jul 28, 2026
Governments are increasingly treating **technology and data sovereignty** as a national security risk factor, weighing dependence on foreign-controlled platforms and supply chains against operational capability. Switzerland ended its use of **Palantir** not over performance, but over residual sovereignty concerns tied to proprietary opacity, foreign legal jurisdiction, and remote update/control mechanisms that could enable remote access, unintended exposure, or service disruption during geopolitical crises. In parallel, U.S. policy discussions are framing “**sovereign AI**” as a strategic export and partnership model, even as partners pursue sovereignty specifically to reduce reliance on the United States amid concerns about shifting rules, access restrictions, and leverage. Separately, reporting on potential U.S. moves to ease certain China-tech restrictions (including around Chinese telecoms and consumer networking products) underscores how quickly policy can change and how those shifts can reshape risk postures for critical infrastructure and technology procurement decisions.
Jul 12, 2026
Organizations are facing unprecedented complexity in securing artificial intelligence (AI) systems as they integrate across multi-cloud environments and even extend into space-based architectures. The proliferation of AI capabilities within major Software-as-a-Service (SaaS) platforms has led to a surge in interconnectivity, with enterprise data now distributed across a patchwork of clouds, databases, and SaaS tools. This interconnected landscape introduces significant data governance and security risks, as each platform has its own unique configuration, visibility, and access control mechanisms. The challenge is compounded by the fact that AI workloads require new types of data movement, access models, and identity management, which traditional multi-cloud strategies are ill-equipped to handle. Data governance becomes central, as organizations must maintain strong classification, control, and visibility over data that is replicated and accessed across multiple AI platforms. Inconsistent policies and roles across different environments make it difficult to enforce uniform security standards. The risk is further heightened by the need for comprehensive user data deletion, as residual data from inactive or deleted users can persist in various systems, including CRMs, analytics tools, and collaboration platforms. If such data is inadvertently used in AI training or inference pipelines, it can lead to compliance violations and reputational damage. Regulatory requirements such as the EU’s right to be forgotten and California's Do Not Share or Sell laws are increasing the pressure on organizations to implement robust data deletion processes. Meanwhile, the expansion of AI security concerns into space-based systems introduces new architectural questions. Commercial satellite constellations now rely on AI to automate security, detect anomalies, and recommend countermeasures, but must decide whether to centralize AI control on the ground or distribute it across satellites. Centralized models offer powerful training but suffer from latency, while distributed and federated models improve response times and privacy but introduce synchronization and resource challenges. The convergence of these issues highlights the urgent need for organizations to rethink their AI security architectures, ensuring consistent governance, robust identity and data management, and adaptable security models that can operate effectively across both terrestrial and extraterrestrial environments. As AI becomes more deeply embedded in critical infrastructure, the stakes for securing data, identities, and systems across diverse platforms have never been higher. Security leaders must prioritize not only access control but also the lifecycle management of user data to mitigate emerging risks. The complexity of managing security across such varied environments demands new frameworks and tools capable of providing visibility, control, and rapid response. Failure to address these challenges could expose organizations to regulatory penalties, data breaches, and operational disruptions. The evolving landscape requires a holistic approach to AI security that spans cloud, SaaS, and space-based assets, integrating technical, operational, and compliance considerations. Only by addressing these multifaceted risks can organizations fully realize the benefits of AI while safeguarding their most valuable assets.
Jun 29, 2026