Being an Innovator at the I7 supporting the G7 summit on Innovation, Industry Science and Labour
I was very flattered to be invited by Sir Mark Walport (the UK Focal Point) to be one of five UK Innovators to attend the #G7InnovationWeek. The five of us: Francine Bennett , Helen Margitts, Jerome Presenti and Andrew Blake arrived in Italy on the Sunday night accompanying Sir Mark and began meeting our fellow innovators in the fields of Big Data, Artificial Intelligence and the Future of Work.
This was one of the best parts of the event from meeting Dyan Regan from Houston who uses drone imagery to support the oil and gas industry in Houston, Sara who has trained 35,000 women and girls in coding in Canada, and Jose who has established a peer-to-peer lending and investment system in Portugal. It was soon revealed that our aim was to provide a communiqué (an official statement) and supporting materials for the G7 Industry Ministers meeting which would run that evening and the next day. But rather more worryingly was that our role was also up for automation and that an AI had been trained on Wikipedia and 20 tops books on AI to prepare an AI version of our communiqué.
Onto the event itself – after a police-escorted bus ride to the event-space (anti-globalisation protesters were due to show up). Once there, we were split into three groups (1) looking at AI (2) on big data and (3) on future of work. Joining the Big Data group we set to work in the morning to prepare our vision for the future of Big Data and in the afternoon to set out the major steps on how to achieve this globally. The results will be published on Friday and you can judge for yourself how well we rose to the challenge of providing something which covers public, private and non-profit sectors at a global level. This supports the main text prepared by the focal points at the same time.
What were my take-aways for ONS and the GSS from this?
- Compared to many others thinking and developing on Big Data and AI, we have a tremendously rich amount of data available to us already so as we develop our Data Science Capability both in the Data Science Campus and in our teams, we should be able to really drive insights which will bring the policy debate to life.
- Text-based AI has really progressed quickly and we need to think whether some of our challenges are best met through text-mining even in our traditional statistical products eg social media mining for online hate crime and cyber-bullying. Many ethical issues but much to explore.
- The personalisation of services and products is going to continue and our forays into online crime calculators and tools which help people interact in a personal manner with our statistics will be incredibly important and powerful.
- The debate between use and privacy is ongoing and lively but was perhaps most perceptively cut through by Kirk Borne from the US who said we could all benefit in the data world from adopting the Belmont principles of ethical research:
- Respect for persons
- Beneficence – do no harm
- Justice (fairness)
I personally thought there was potentially a lot of power from these simple principles.
- The need for ethics in AI was well-recognised and understood to ensure that AI lead to public good and are non-discriminatory. Ideas on how to do this were less forthcoming, with the main one being a body similar to the Nuffield Centre on Bio-medical research.
- The Justice DataLab is not just known in the UK – it is known internationally with Lesley Cheung from Canada raving about it to me and how good it was.
- The changing world of work. A strong desire to know and understand which jobs were being impacted by automation – both growth and declining and a true understanding of the gig economy. I could be wrong but I suspect this may make it into the main text. I also think the focus on lifelong learning will only increase which will mean DfE’s skills data will be even more important.
We finished with our work being presented to the G7 Ministers alongside the reveal of the AI generated communiqué. The photos below show the results from the AI – and I was pleasantly surprised by it.
It was then speeches from Ministers and a reception with them before we went back to the hotel after a long day to clear my e-mail backlog and get ready for an early start and flight back to London on Tuesday morning for Government Statistical Service Heads of Profession meeting.
We’ll know later in the week how much impact we had but for me one of the main benefits was seeing and hearing from all the others there at the event and what they were already achieving – making the scale of our task all the clearer and leaving me enthused about the possibilities.