Lifebit launches AI Scientist Studio for governed health data

Lifebit launched AI Scientist Studio at HumanX Amsterdam on September 24, 2026, giving researchers a way to build custom AI scientist agents inside federated Trusted Research Environments. The platform is designed to let frontier AI run on sensitive health and genomic data without moving patient-level data out of governed environments.

Why it matters: - The launch aims to close a major gap in biomedical research: frontier AI tools are advancing quickly, but sensitive health data often cannot leave controlled environments. - Lifebit is pitching a model where AI moves to the data, rather than forcing governed datasets into third-party systems. - The approach could help governments, health systems, biobanks and life sciences organizations use advanced AI while keeping control, sovereignty and oversight intact.

What happened: - Lifebit announced AI Scientist Studio at HumanX Amsterdam on September 24, 2026. - The platform lets researchers build and deploy custom AI scientist agents inside secure Federated Trusted Research Environments. - Lifebit said the system supports frontier AI models on governed health data without requiring patient-level data to leave the source environment. - Dr Maria Dunford, Lifebit CEO, said the product changes the model by bringing any AI model to the data and letting researchers run AI scientists securely across health and genomic data.

The details: - AI Scientist Studio is built on Lifebit’s Federated TRE architecture. - The platform applies the Five Safes framework to agentic AI. - Researchers can bring their own foundation model, or choose one, along with scientific tools and knowledge bases. - Users can build specialist agents for cohort discovery, deep phenotyping, analysis and evidence generation. - Agents run on compute inside each data custodian’s governed environment and inherit user permissions. - The platform operates federated across sites. - Patient-level data stays in place. - Chat history, outputs and full audit trails remain within the environment. - Only approved results leave through Lifebit’s Airlock controls. - Lifebit said the platform is model-agnostic.

Between the lines: - The launch reflects a broader tension in health AI: research teams want frontier models, but privacy, compliance and data-sovereignty rules limit how far sensitive data can travel. - Lifebit is positioning AI Scientist Studio as an infrastructure answer to that problem, not just another analysis tool. - The company is also betting that customizable AI agents will be more useful than fixed, pre-built tools that researchers cannot easily adapt. - Dr Dunford said major AI vendors have moved into scientific research this year, but the dominant approach still assumes data comes to the AI.

What's next: - Dr Maria Dunford was scheduled to discuss AI Scientist Studio at HumanX Amsterdam between September 22 and 24, 2026. - Lifebit plans to focus on how organizations can apply frontier AI inside governed health-data environments. - The company will likely lean on its existing federated infrastructure as it rolls out the new studio. - Lifebit said its platform already supports secure management and analysis of more than 275 million patient records across 30+ countries. - Existing users include UK’s Genomics England, UK’s NIHR BioResource, Canada’s Canpath, Singapore’s Ministry of Health and Boehringer Ingelheim.

The bottom line: - Lifebit is trying to make frontier AI usable for some of the world’s most restricted biomedical datasets without weakening data governance.