Guest blog by Jonathan Everett, RSS Head of Policy
It has been another busy three months: the UK has got a new PM and a new set of ministers, changing some of the priorities of the government and causing us to reflect on how statistics, data and evidence can help deliver those priorities. We’ve also seen the announcement of a new National Statistician, who we’re looking forward to working with when he is in post.
We’ve also continued our work on AI; thinking about how the technology is evaluated and regulated has been a major focus for us over the past quarter.
Our ongoing work on AI continues to spark conversation
Earlier this year we released a position paper, AI is Statistics, that made the case that AI is a fundamentally statistical technology and that this matters for how we interact with and evaluate it. In July, we built on this with the release of AI Regulation Needs Statistics. This paper argues that effectively regulating AI is in large part a question of effectively evaluating AI models as they are operating in a range of settings – and this is a statistical challenge. The paper looks at three case studies – health care data, education and financial advice – and draws some key lessons for effective regulation.
Since some of these lessons relate to how regulators incorporate statistical expertise into their work, we have reached out to the regulators of the sectors discussed. We have had warm responses from a number of these so far and have meetings lined up to discuss how they might make use of the expertise of our members to inform their work.
We have seen a really significant response to our work in this area from members and the wider community, and it has been fantastic to feel the conversation moving forward.
Elsewhere, our AI Task Force was busy at the RSS Conference, running a number of sessions to update the wider community on our work. We’re planning an event for Members Week in October to discuss our work around regulation of AI, so keep an eye out for that if you’re interested.
We also submitted a consultation response to the government's call for evidence on data regulation in the age of AI. Our response focused on: issues around anonymisation and the increasing risk of re-identification arising from AI use; accountability, strong governance and transparency around AI-generated uncertainty; and the need for safeguards where AI informs decisions.
Public statistics: consultation responses and engagement
One of our major concerns in the realm of government statistics was the amount of time that it was taking to appoint a new National Statistician. We urged the new government to resolve this as quickly as possible and so were pleased to see the news that André Loranger will be taking up the post later this year. We’re looking forward to working with him in due course.
The Public Administration and Constitutional Affairs Select Committee has also recently released its long-awaited report on the state of the official statistics system. Our Public Statistics Advisory Group is preparing a response, which we’re aiming to release in October to feed into the lessons government learns.
We have continued to engage constructively with the statistics system. In July we held a roundtable with the ONS to bring together statistics users, producers and regulators to discuss coherence and comparability across UK statistics. Our Honorary Officer for Public Statistics, David Caplan, has reflected on how the discussion highlighted both the complexity of the challenge and the strong appetite for greater coherence and comparability across the UK statistical system. We plan to continue to engage with the ONS as they develop their strategy on coherence and comparability.
Finally, we responded to the government’s consultation on its proposals to charge for the re-use of public sector data. We argued that public sector data should remain open and accessible wherever possible. We were invited to meet with the team at DCMS who are developing these proposals to discuss our response in person.
We have also continued to draw attention to the potential issues with the government’s proposal to introduce pay gap reporting for ethnicity and disability. We are concerned that the plans as they currently stand could lead to unintended outcomes, such as incentivising firms to hire fewer new people with disabilities in order control a pay gap. We wrote to the Equalities Minister to make this case and will continue to push to ensure that this is recognised as plans develop.
Public Understanding: developing crucial resources for specific groups
Excitingly, this year’s cohort of RSS William Guy Lecturers have begun their terms. You can watch their specially developed talks, each designed to educate a particular age-range about the theme of Statistics and Misinformation, online now. Our three lecturers – Tom Crick, Jonathan Davies and Takoua Jendoubi – have all produced interesting and engaging talks and I’d encourage everyone to give them a look and share them with schools and maths teachers in your networks.
Meanwhile, we are pleased that our efforts to engage with exam boards to help them think about how they assess statistics (on the grounds that teachers are likely to adapt their teaching in response to how it is assessed) have met with success: members were invited to join WJEC’s Qualification Development Advisory Group for their V Level Digital Systems and Data qualification. This is something we hope to expand to other exam boards and qualifications. We have also been preparing for the publication of the new mathematics curriculum by establishing principles of good assessment, which we plan to publish soon.
Finally, we continue to add to our series of explainers, created to illuminate specific areas of statistical relevance. Since last writing to you, we have published a new set of FAQs on Statistics, Misinformation and AI, developed by members of the RSS Education Policy & Public Engagement Advisory Group. The resource explains key concepts in accessible language and offers practical advice on using AI critically and responsibly. Another explainer from the RSS Climate Change Task Force examines one of the most important questions in climate policy: has the 2015 Paris Climate Agreement actually worked? It sets out that, despite small improvements, progress is not on track to meet the objectives of the Paris Agreement, with only a 17% chance of keeping warming below 2°C. But, if countries fulfil their existing commitments under the agreement to 2030, and then continue efforts at the same level, warming could be limited to 2.1°C. (That feels like quite a big hypothetical – but at least some grounds for hope.)
Pictured: Jonathan Everett with members of the RSS AI Task Force at the RSS International Conferece 2026 in Bournemouth. Photo @rugfoot photo