Cookies on the Gambling Commission website

The Gambling Commission website uses cookies to make the site work better for you. Some of these cookies are essential to how the site functions and others are optional. Optional cookies help us remember your settings, measure your use of the site and personalise how we communicate with you. Any data collected is anonymised and we do not set optional cookies unless you consent.

Set cookie preferences

You've accepted all cookies. You can change your cookie settings at any time.

Skip to main content
  1. Home
  2. About us
  3. Publication scheme

Panel minutes for 17 January 2024

17 January 2024, 10:00 to 11:30 Virtual meeting via Microsoft Teams

Panel Members:

  • Andy Payne, (Chair)
  • Francine Bennett
  • Simo Dragicevic
  • Steve Homan
  • David Hurst
  • Paul Smith
  • Alison Pritchard
  • Darren Williams.

In Attendance:

  • REDACTED (item 4 only)
  • REDACTED (up to 10:45 only)
  • REDACTED,
  • REDACTED (minutes),
  • REDACTED
  • REDACTED

Apologies: Christian McMahon.

1. Welcome and introductions, apologies, note of last meeting and declarations of interest

DAP welcomed everyone to the meeting. The new Head of Expert Groups and Governance Stakeholder Manager introduced themselves to DAP.

Apologies were noted from Christian McMahon.

There were no new declarations of interest.

The note of the meeting held on 10 November 2023 was approved as a true record of the meeting.

2. Update on data programme

DAP noted the papers.

DAP were given a brief verbal update on the key work completed on the data programme since the last meeting.

Sub-group meetings are being arranged for late January and February to obtain further input to our work on pilot project planning, development of the Regular Feed of Operator Core Data (ROCD) project and capacity building.

On capacity building, 2 data scientists have been recruited who will be able to take on more projects. There remains a data engineering gap which the sub-group will pick upon the ROCD project, we went through process of getting legal advice and have worked the implications through into our data spec. We can soon start engaging with operators in the pilot phase and are in good shape to launch into the next phase of that work. There are still some questions for the sub-group to advise on.

On pilot projects, the last quarter has been particularly enjoyable delivering some of these projects and we are now in a position to look at what we can scale up and which new projects we can take on. There are some opportunities to build some products for the compliance team and project prioritisation is being carried which we will sense check with the sub-group.

The reception within the Commission on the data work has been positive and we have been doing some internal comms and have joined some team meetings, asking colleagues what issues they have that could be solved by data. The Executive team also received a presentation at its last meeting, as did the Board in September. The data project is also being built into the Corporate Strategy for the next three years so visibility within the organisation is growing. The Commission Chair has also been receiving regular updates and is very supportive of what is being done. In addition, we briefed three other Commissioners on the data work on Monday and have a Teams ‘outputs of the data project’ session with Board and the Executive team on 13 March.

DAP commended the team on the really good progress made.

3. Illegal gambling

DAP noted the papers.

REDACTED advised DAP that work has been ongoing since the summer with Simo Dragicevic to scope metrics and to help us better understand illegal gambling. One of the data projects was to really understand more detail in measuring illegal gambling in the UK, trying to scope out the art of the possible.

Simo talked through the previously circulated slides.

DAP was invited to give views on plans before we undertake further work on implementation:

  • REDACTED
  • it would be worth doing a literature scan or review out of academia on illegal markets. Paul can follow up on that and will send a list of reports to Simo
  • on the question of whether this work is a pilot or a project, the scoping that has been done is working out what the options are and we would like to see three to six months more work, which would make it a pilot, but if the work were to go beyond that it would move into the scope of a project.

DAP members were asked to contact Simo and the Commission team with any further thoughts following today’s presentation.

4. Discussion on AI – GC approach and risks to be managed

DAP noted the paper.

DAP received a brief summary of what we would like to seek DAP’s input on to inform work on this topic for the next quarter. This will include:

  • developing plans for a thematic project on AI where we will ask a selection of operators to tell us how they are using it and how they are managing the risks
  • preparing for a panel session on AI at our spring conference. This provides an opportunity to highlight the types of risk we think operators should be thinking about
  • speaking to our Audit and Risk Committee (ARC) about our views on risks associated with AI deployment in the gambling industry.

It would be useful to get DAP’s advice on key risks and opportunities in relation to AI so that we can consider how best to highlight them through the activities above.

The key things we are starting to do some thematic work on are included in a plan that has been shared with DAP members, who are asked for views on risks of how AI is used. We have the Commission’s annual conference taking place in March and there will be a panel to talk about AI deployment, so it would be helpful to get views from DAP on what are the issues we should be thinking about. Finally, the topic of how AI risks are being approached in the Commission will be discussed at the next ARC meeting on 1 February.

The DAP made comments and asked questions.

AI should be considered an overarching term – we should endeavour to start the conversation on what the operators are likely to be using AI for. AI should be categorised – all categories have different connotations - rather than being used as a generic, AI term.

In a wider context, central Government has really gone back and forth on how they want to regulate generative AI and are being quite hands off at the moment REDACTED. Helpful to consider the issue alongside the resourcing for the Commission and where we might need additional resources in the future.

There are large global services organisations who are in partnership with Microsoft and they are using the large language model as a front end – a comms tool – and are being very careful and very structured in terms of just using proprietary data. You can do quite a lot with that without huge amounts of data so we might see something like that quite quickly – an interface – it is getting cheaper – and there are options outside of the big players (article shared on AI risk:).

Gambling companies are not going to create new models, but we are seeing a lot of fine-tuned models. It will only take one operator to set off others.

The big AI costs are in training the models, the tuning of the models is comparatively cheap and it would not be surprising if some operators are not already using those methods.

There are 2 distinct lenses here: segmentation around user cases does matter and, in that regard, whilst the main focus is on generative AI and large language models, we should also be looking at other areas of machine learning in the sector. In addition, we should have a good, strong lens on the other tools and techniques in AI too.

Sports betting, particularly in the US, is rapidly being legalised, around automatic generation of odds and content. The hype is how do you use AI in different forms to super personalise sports betting which is very big in the US, and it would not be a surprise if the UK were soon to follow. Noted that data is already used extensively to profile players and customise marketing and product offerings.

There is a lot of talk and hype around accelerating what you can do in the gaming industry, where there has been some ‘de-hiring’ as the industry has been using machine learning for a number of years. This is now moving very fast in the games industry and we can see a lot more crossover with ‘games’ and ‘gaming’ for example, loot boxes.

Profiling and tailoring customers is well established, more of a step change to a different type of interaction – becomes much more ‘human’ like but is not – and the problem is the scaling up of that intervention.

Worth considering focus on AI as an opportunity to achieve greater transparency from operators about how their risk detection and customer profiling works. There is already a lack of transparency around use of algorithms – the focus on AI might be an opportunity to revisit this.

Human like chat bots is definitely a good area to be explored. It would also be good to use this as an opportunity to get under the bonnet of what is being done with machine learning and to have more transparency and disclosure in the areas that are being used in consumer protection, etc.

The DAP’s attention was drawn to an article in December 2023 which looked at using large language models with smaller data sets to pinpoint problem gamblers (opens in new tab) (PDF).

REDACTED.

5. Any other business

There was no other business to discuss.

Date of next meeting: 27 February 2024

Is this page useful?
Back to top