IBM и Duality Technologies объединили технологии суверенитета и приватности для работы ИИ

- IBM and Duality Technologies are combining sovereign controls with privacy technology - The approach is intended to let AI workloads operate without pooling sensitive source data - Organizations can retain control of access, encryption keys and operational records IBM Sovereign Core and Duality Technologies have partnered to integrate sovereign software controls with privacy-preserving technologies, enabling artificial intelligence and analytics in distributed data environments while maintaining control of sensitive information, IBM said Thursday. How Does the IBM-Duality Collaboration Support Sovereign AI? Duality allows approved analytics and AI workloads to interact with protected data within organizations without centralizing or pooling raw information. Meanwhile, IBM Sovereign Core maintains identity, access controls, encryption keys, operational logs and continuous compliance evidence within the customer’s sovereign environment. IBM noted the approach addresses operational challenges for public-sector agencies, defense missions, healthcare systems and financial institutions, where regulatory and sovereignty rules prevent sensitive data from leaving its original jurisdiction. IBM introduced Sovereign Core as software designed to help organizations maintain control over AI-ready environments, including deployment decisions, system operations, operational data and audit trails. The software provides the sovereign control layer for the new IBM-Duality integration. What Capabilities Does the Combination Enable? IBM outlined three use cases for the integrated capabilities. The first is federated workloads, which let approved AI and analytics run across departments and sovereign clouds without pooling raw data. The second is protected use of public cloud capacity, which extends approved workloads to scalable resources while maintaining security and governance. The third is confidential use of external data sources, where sensitive queries remain protected when interacting with outside data providers. According to IBM, the offering lets organizations run vetted workloads across agencies while collaborating without exposing source data. Agencies can also tap public cloud resources without losing control over outputs and generate combined insights without standing up a new centralized repository.