Introducing the AIR Platform
RegulAItion’s privacy-preserving data access network for regulators and regulated industries
In July 2020 data platform company RegulAItion announced it had been awarded a Government backed grant from UK Research and Innovation to lead the development of a breakthrough data access project - the AIR Platform - alongside its own hand-picked group of Tier-1 partners and collaborators. The AIR Platform is a pioneering data platform for privacy-preserving data collaboration between regulators, regulated industries and their professional services providers such as lawyers and accountants.
Collaborators joining the project include the FCA, two international banks, Ashurst, Wilson Wright and Oasis Loss Modelling Framework as well as academic partners University College London (UCL) and Loughborough University (LU).
AIR Platform as an infrastructure will be completed in Q3/2020 with use cases being concurrently developed with collaborators. Full completion with real-world return on investment results is scheduled for Q2/2021.
AIR Platform: What is the problem?
There is simply more data in the world than businesses can handle, and it is suffocating industries and regulators. Equally, data silos mean organisations are unable to develop meaningful solutions, and privacy concerns such as GDPR and commercial interests stand in the way of delivering collaborative efforts to share knowledge from data.
As a result, the public and private sectors face two significant challenges:
- Data management: how to acquire, validate, store, and process required data to ensure the accessibility, reliability, and timeliness of the data for its users - which includes machine learning technologies specifically for this purpose
- Data collaboration: providing access to sensitive data while preserving privacy. This is especially important where it is necessary to train machine-learning algorithms with sensitive data across multiple institutions.
It is widely accepted that AI is critical to enable any form of fundamental breakthrough in these areas but without access to ample, relevant and good quality data, AI technologies cannot develop. Data-holders need trust and assurance to be confident in providing access to their data. And some of the areas where data is most sensitive - commercially or personally - are also the ones where the greatest benefits lie.
This is especially true in the regulated sectors, such as financial services, including banking, insurance and asset management, where legally mandated Chinese walls prevent data from being shared even internally within the same organisation. In legal services, legal privilege and conflict management compound the problem. In addition, even when data-holders see a case to proceed with secure data access, and trust the external organisation they plan to collaborate and transact with, they often lack the essential resources of time and money, expertise and know-how to form agreements, establish trust between the parties and manage the data access processes in practice.
“Tackling these challenges as individual businesses or sectors is not only prohibitive but also limiting in terms of what can be achieved that is of real value. However, a collaborative, sector-agnostic approach driven by artificial intelligence, machine learning and blockchain technologies can work.”
Sally Sfeir-Tait, CEO, RegulAItion.
AIR Platform: Creating the solution
The solution comes as a privacy preserving, data access and data collaboration platform that will deliver a ‘generational breakthrough’ to how the private and public sectors operate and interact with each other locally and globally.
Using a technology stack that preserves security, privacy, regulatory and commercial protections in an automated way, the AIR Platform allows third-party managed data access and collaboration using privacy enhancing techniques including federated learning. This platform will offer its users - data-holders and AI providers - security, privacy, regulatory and commercial protections in an automated way on pre-approved standard terms.
In 2017, an independent review into AI, commissioned by the UK government, recommended data trusts to share data in a ‘fair, safe and equitable way'. Federated learning is an application of data trusts in the form of a data alliance between a group of alliance members for a specified purpose. By bringing the algorithm to the data itself, federated learning solves the fundamental issue of privacy, as data never leaves the secure premises of the original data-holder. As a result, the AIR Platform will enable any form of data trust relationship to be implemented be that bilateral (e.g. between a law firm and an AI start-up) or multilateral (e.g. between two insurance companies and an insurance broker).
Bringing immeasurable benefits, data holders now only complete a ‘connection exercise’ once. After that, they can decide who to provide data access to and for what purpose. This access can also be revoked in an automated way. This is a significant leap forward. Systematically providing data access through a trusted platform will help create a competitive and innovative environment that serves the interests of regulators, holders of large datasets who want to develop next generation services, and of companies who could develop AI services by leveraging that data.
“Our vision for the AIR Platform is to provide the digital infrastructure required for scalable, automated, repeatable, and responsible data-access, supporting the Fourth Industrial Revolution and the UK's leading position in it.” Sally Sfeir-Tait, CEO, RegulAItion.
For more information about RegulAItion or to contact the team go to www.regulaition.com