Enabling rapid adoption of artificial intelligence through an anonymized data protocol and explainable models

Two of the biggest barriers to using artificial intelligence in UK services are the need for confidential data and the fact that many AI models are black boxes that cannot explain their decisions. This project brings together academics, large companies, and a machine learning startup called Ginie AI to solve both problems at once. The researchers will develop commercial products that advance the state of the art in computational privacy and machine learning. They will test these technologies in real business settings. At the same time, they will work with regulators and industry stakeholders to create a standard protocol for sharing data safely, so that organisations can adopt AI more quickly without compromising privacy. If this succeeds, services that rely on sensitive data—such as healthcare, finance, or public administration—could deploy AI systems that are both secure and transparent. Instead of a black box that gives an answer without explanation, a hospital or bank could use a model that explains its reasoning, while the underlying data remains protected. The project does not aim to make fundamental scientific discoveries; it is focused on removing practical roadblocks so that existing AI research can be put to use at scale.