Tim Hubbard (ELIXIR) - Navigating the Future of Secure AI and Health Data within TREs

Описание: Tim Hubbard, Director of ELIXIR and Professor of Bioinformatics, discusses the challenges and opportunities in managing biological and health data within Trusted Research Environments (TREs) and the evolving infrastructure supporting these efforts. He focuses on European initiatives such as the European Health Data Space and the Federated Genome Data Infrastructure (GDI), which aim to integrate clinical and research genomes into secure systems for large-scale data analysis, demonstrating the potential for combining research and clinical data to drive scientific discovery while maintaining privacy and security.

Professor Hubbard highlights the increasing sensitivity of genomic data and the critical need for secure data environments. He emphasises the importance of the five safes framework (safe data, safe setting, safe people, safe projects, and safe outputs) to ensure that data can be used securely for research, particularly when training AI models. The challenge lies in ensuring that AI models do not inadvertently capture private information, with ongoing research exploring whether AI models can be trusted to avoid such risks.