Secure AI Labs
Secure AI Labs offers a secure, easy-to-manage & auditable solution for hospitals to track & trace the usage of their patient data through clinical research.
Using the Secure AI Labs platform, data never leaves hospital firewalls while enabling scalable collaboration. Hospitals keep the data safe, and researchers can innovate faster than ever. Healthcare data sharing enables collaboration and innovation; unfortunately, the existing ways of data sharing at hospitals are unscalable and insecure. Hospitals need a solution to manage access, track and trace usage of data, and maintain full control throughout the entire research process. Using SAIL, hospitals can track, trace, manage access, and collaborate without compromising data security and privacy. Patient records remain fully within the firewalls of the hospitals, and hospitals are able to revoke sharing rights at any time to any parties.
The next wave of healthcare innovation will depend on information fused and analyzed from electronic medical records, genetic biobanks, clinical trial data, insurance companies, government agencies and other research organizations. Traditionally these data sources are protected through technical transformations (e.g. anonymization) and extensive regulatory, compliance and legal processes. These safety measures are critical to minimize disclosure of personal identifiable information. However, these processes add significant cost and add lengthy delays to data research, limiting overall research capacity and effectiveness.
SAIL was founded by a team out of MIT to radically improve the opportunity and efficiency for collaborative, secure data research. They are building a platform to enable research of health records across platforms in a regulatory-compliant manner. Their initial partner and customer sector focuses are healthcare research organizations which need more efficient ways to enable research across data categories and to access comprehensive health care records in their R&D efforts.
When researchers wish to access healthcare data it can often take 6-12+ months for the data to be prepared, anonymized and packaged for compliant sharing. Often the burden of this data aggregation makes it untenable to move forward with a desired research investigation due to timing, cost and/or potential liability.
SAIL flips the model. With the SAIL solution, the data never leaves the custody of the provider’s data warehouses, and instead researchers securely push machine learning models into a distributed, secure and remotely managed computing environment. The system leverages a technique called "federated learning", which allows the researcher's AI algorithms to be shared across multiple providers’ data warehouses, the results of which then get unified back in the SAIL platform. The SAIL platform is built upon 3 core components: Federated Learning protects data by computing analyses at the data source; Secure Enclaves protect the remote analyses by encrypting the computation; and immutable audit records for regulatory and legal compliance.