Artificial Intelligence Regulation

Posted : January 2, 2024 -


On the 9th Dec 2023, following a marathon 37-hour negotiation between the European Parliament (EU) and the EU member states, agreement on the EU AI Act was finally reached, putting the EU ahead of the US, China and the UK in the race to regulate artificial intelligence. Meanwhile UK’s Artificial Intelligence (Regulation) Bill was introduced to the House of Lords on 22nd Nov 2023 for the 1st reading Artificial Intelligence (Regulation) Bill [HL] – Parliamentary Bills – UK Parliament.

You don’t have to be a Chief Data Officer or a Chief Technology Officer to have an opinion on the regulation of AI in the UK. As a body of analysts foremost, we should all be aware, use our moral and ethical compass and form an opinion in the use and implementation of AI in our own organisations.

The UK’s Artificial Intelligence (Regulation) Bill is 4 pages long (I was expecting something with more content when I first came across the bill). However, the content I do believe is setting the building blocks for regulation to actually work; as well as setting out the Regulatory principles (the standard ones you would expect), the creation of a Government AI Regulatory Body, creation of a Sandbox for innovation and testing and lastly the requirement to appoint an AI Officer in any business or organisation that deploys AI. For starters, I never thought I would read the word “Sandbox” in a Bill!

Having worked at University Hospital’s Birmingham (UHB) NHS Trust; I was aware that as an organisation, there was a real interest in exploring how AI technologies could be used in quicker diagnosis and UHB was working in partnership with University of Birmingham, the NHSX NHA AI Lab, the National Institute for Health Research (NIHR) and the Health Foundation Project uses AI to improve AI systems (uhb.nhs.uk) from 2018. It was great to see UHB get stuck right into this in the same way it leads on implementing digital technologies.

After 3 ½ years at UHB, I joined the Civil Service (Ministry of Justice – HMCTS) in April 2023; I could “hear” there was an increasing interest in using Generative AI Tools to improve, how we worked as individuals, the User Experiences of HMCTS and the efficient running of Court and Tribunals in England and Wales. It was to my surprise that this was being thought about as early as 2018 16808e4d87 (coe.int). In 2023, the following guidance on the use of generative AI by Civil Servants, was published by the Cabinet Office (Central Digital & Data Office) Guidance to civil servants on use of generative AI – GOV.UK (www.gov.uk).

Perhaps we can then helpfully “organise” AI uses and regulation into three scopes “Individual”, “Organisational”, “Customers/Users”:

 

The requirement for a designated AI Officer role naturally makes you think how would this be best implemented? Should it be added to the Chief Data Officer or Chief Analytical Officer responsibilities or regarded as an “independent protected role” like the Caldicott Guardian UK Caldicott Guardian Council – GOV.UK (www.gov.uk)?

As analysts what is your role? I believe we can take small but useful steps like agree on a recommended list of off-the shelf AI tools we can use (and how we use them) to do our work better, challenge (ask politely) how any new AI tools being introduced to your department or organisation are trained to ensure they are not biased and existing processes are updated so they do not inadvertently exclude people from fully benefitting from AI. For example this advancement AI to speed up lung cancer diagnosis deployed in NHS hospitals – GOV.UK (www.gov.uk) was really interesting to read but my initial thought was that the current process relies on a GP to refer the patient to a hospital for investigation. By the time the patient becomes aware there is something “wrong”, goes to see their GP, the GP makes a referral and the patient is seen at the hospital; too often than not, hidden cancers such as lung cancer are diagnosed at Stage 4, when it is too late to do anything.

We are in the dawn of new age and one way or the other, AI is coming and will change our lives but we have to ensure we look at implementation holistically to realise the full benefits of AI for individuals and the wider society.

Written by Chandan Kaur, Head of Reporting & Analytics, Data Analysis and Insight at HMCTS.