In December 2019, RegulAItion was informed it had been awarded a grant from UK Research and Innovation to develop, test and launch a pioneering privacy preserving, data access and data collaboration platform - the AIR Platform (Automated Intelligent Regulation). Co-Founder and CEO of RegulAItion, Sally Sfeir-Tait, talks through the significance of the UK Research and Innovation event and what the subsequent grant has meant for the company.
How did RegulAItion first become involved with UK Research and Innovation?
You cannot be an innovative business without knowing about UK Research and Innovation. It was quite simple really, I’m on their mailing list, saw the Innovation Lab event and its focus on data access and knew I wanted to be there. RegulAItion’s Chairman and UCL’s Director UK Centre for Financial Computing, Philip Treleaven also contacted me to suggest I apply to take part. Data access is the single biggest problem holding back the development of AI. This is a problem that we as an organisation want to solve so of course we were going to get involved. I had to apply.
It was a rigorous two-phase selection process to be included in the programme. In the first round they invite individuals – based on ‘the individual’ and not the organisation they represent. Phase 1 success is to secure an invitation to the Innovation Lab and create a viable project. Phase 2 is about progressing this project and applying for grant funding.
I was in our London office in October when I was informed I’d been selected for the Innovation Lab. I remember it well. We had just finished an OKR (Objective and Key Results) session. I was super excited by the opportunity, but I could never have imagined how instrumental it was going to be to the future of RegulAItion.
Can you explain more about how the event was run and how you reached your ‘breakthrough’ moment?
On the first day I wanted to leave. On the second day I almost left. It was brutal. It was brilliant.
It was a three-day intensive workshop with around 50 attendees. On day one, after a 5am start from home, we had a detailed programme introduction and were put into groups. Then the real work started. Each day entailed more than 12 hours of intensive discussion, debate and solution creation. On day two my group began at 7:30am and finished work at midnight! I wish I had recorded my impressions at the time as there was so much that happened!
On the second day, now in our groups, the team from UK Research and Innovation steered us through the programme to identify the problem, the solution, the market etc. There’s a specific format we all had to follow. And during that morning our group’s first idea was slammed. But one of the design-thinking coaches took me aside and told me I was on to something and not to give up. Seriously, they were so good. That challenge, the frustration I felt along with the encouragement I received propelled me forward to think even deeper and wider around the data access issue to get to the breakthrough moment.
What was the breakthrough that led you to defining the AIR Platform?
Federated learning. On the morning of the second day federated learning had not come into the conversation, I was still focused on ‘permissioned access’ based on the blockchain system that we had developed with the FCA and the Bank of England. But, during the evening, while preparing for the big presentation we went deeper in our research. By 11pm the penny had dropped; the breakthrough we had been looking for lay in the combination of algorithms and blockchain technologies. I called my co-founder and Head of R&D, Hirsh Pithadia and overnight he developed the conceptual architecture of what is now the AIR Platform.
The next day, at 6am, I again called Hirsh to talk everything through and to check that we could deliver the platform. Luckily, our official presentation wasn’t scheduled until midday so that gave us the whole morning to refine our proposal. We expanded permissioned access into federated learning and presented our new idea for a privacy preserving, data access and collaboration platform. And we nailed it.
Once the UK Research and Innovation event was over how was the funding secured?
Phase 2, following on from the Innovation Lab, was to officially apply for funding. We had a very intense four weeks to get the application through. A significant element of this was to confirm the industry collaborators which is a key requirement from UK Research and Innovation. Safak Dogan from Loughborough university had been in my group from the first day and remained a willing partner. UCL then came on board. I knew the industries I wanted as collaborators and I hit the phone. I often talk about ‘Magic Thursday’, which was the day when everyone I had talked to said ‘yes’. The AIR Platform was born, and people have not stopped saying yes ever since.
Shortly after that, around 20th December, we received the news that we’d been successful. We were one of five projects selected by UK Research and Innovation for funding. It was the best early Christmas present.
What did this mean for RegulAItion as a company?
The AIR Platform is a £1.67 million project jointly funded through private investment and the grant. The funding is important but, more than that, it’s the backing of UK Research and Innovation and their validation of a concept that is beneficial to the market and to the UK economy. That’s the most valuable thing for us.
How is development of the AIR Platform progressing?
RegulAItion had already developed multiple elements of the AIR Platform in previous work, including the work done with the FCA and Bank of England. So we had components of the first iteration ready. We’ve developed and completed our internal POC that tests our theory and we’re now in full development. The AIR Platform as an enterprise-grade infrastructure will be completed by this September and we are developing use cases concurrently with our collaborators validating real-world return on investment to create the hard-proof the industry needs to get on board.
Our ultimate delivery deadline from UK Research and Innovation, for the project and the results of the use cases with the collaborators, is the end of June 2021.
What has been the impact of UKRI’s Innovation Lab for RegulAItion?
It crystallised our vision.
Because of my legal background, I had focused on the regulation problem and the need to deliver a platform to enable machine readable and executable rules. I hadn’t yet focused on the data problem or, more specifically, the data access problem. If we are ever going to deliver a straight through process of machine executable regulation – real-time regulation - you must solve the data access issue; it is the cornerstone of the problem. The Innovation Lab helped us identify this missing piece and gave us the environment in which to find the breakthrough.
Would you say Government-led initiatives such as this are important for the technology industry?
Absolutely. But only if you are daring to change the world for the better. It’s all about innovation that has potentially disruptive effects and a positive impact on the UK economy. That is what RegulAItion as a company and the AIR Platform are all about.