Intel works for safer and more responsible AI

Artificial intelligence technologies have started to take place in every aspect of our lives today, and the number of sensitive data collected is increasing day by day. The proliferation of sensitive data is causing the threat landscape to grow and security concerns are increasing.

secure and responsible artificial intelligence from Intel

Intel’s industry-leading research efforts; focuses on protecting the integrity, accuracy and confidentiality of data. In this regard, it is taking two important steps to accelerate AI deployments in responsible and secure ways.

The new solution, called BeeKeeper AI, leverages Intel SGX hardware-based security capabilities and Microsoft Azure’s confidential computing capabilities to provide a zero trust platform. using the infrastructure. It allows an AI algorithm to run on multiple real-world clinical datasets without compromising the privacy of the data or the intellectual property of the algorithm model. This accelerates AI development and deployment innovation in healthcare by more than 30% to 40% compared to the current method.

The second step was to use a suite of Intel hardware and open source software technologies to improve the training of artificial intelligence models to find brain tumors in research partnership with Federated Tumor Segmentation of the University of Pennsylvania Perelman School of Medicine.

Intel technology using Open Federated Learning (OpenFL) enables any organization to contribute to improving the fidelity and quality of inference algorithms. OpenFL enabled 55 institutions across six continents to collaborate while maintaining the security and privacy of individual datasets. This resulted in an AI model that improved efforts to locate tumors by 33%.

Responsible use of AI is also an example of how industry can come together to pave the way for distribution in a variety of industries including healthcare, financial services, manufacturing, retail and entertainment.