Agenda item

This paper provides an update on artificial intelligence (AI) adoption at H&F and on AI use cases in other Local Authorities, and shares H&F’s draft governance framework for AI.

Tina Akpogheneta (Chief Digital Officer) provided an update on artificial intelligence (AI) adoption at Hammersmith & Fulham covering the following points:

• Since the last update to the Board in April, the working group had been extended to include different services.
• Microsoft Copilot for Edge had been enabled so staff could get familiarised with using AI at work.
• A comprehensive AI Governance Framework had been developed.
• The view of the team was that Hammersmith & Fulham would likely benefit from more purpose-built tools and from tools that were integrated into the systems.
• Resident Services was using Copilot for Word to assist with complaints procedures, while Social Care was exploring using AI solutions to generate Education, Health, and Care Plans. Housing was implementing Voicescape for tenancy analytics.
• The recommendation of the team was that Hammersmith & Fulham needed to ensure its data was AI-ready and that the costs of adopting AI should be managed. There would be continued work on data literacy and data stewardship among staff, and options on managing costs would be taken to the Council’s Senior Leadership Team (SLT) shortly.

Sam Nutt (Researcher & Data Ethicist at the London Office of Technology and Innovation) gave a presentation covering how other London councils were using AI and the public attitude towards AI. A few key points for adopting AI responsibly were highlighted:

• Start with common ethical principles
• Ensure compliance with common legislation (GDPR, Equalities Duty, Human Rights)
• Embed your standards and policies into procurement processes
• Invest in proper evaluation of AI uses
• Engage with residents and staff to understand concerns better

Councillor Natalia Perez enquired about what Hammersmith & Fulham was doing to nurture public trust in the Council’s use of AI and the role of co-production in the process. Tina Akpogheneta said the use cases were currently small scale, but the team would look at options to build public trust when they started to scale up to larger use cases. It was noted that co-production and training on ethics for decision-makers were important going forward. In the meantime, AI projects would go through the Digital Board, which was comprised of SLT members and others.

Councillor Jacolyn Daly was interested in how AI was being used in the Housing department. She applauded the slower move around using AI for any process that could penalise residents. She was interested in how AI could be used in budgeting, predicting demand, and making information more accessible to residents as well as the use of data for predictive analytics. Tina Akpogheneta said that at the moment, the use of AI in Housing was limited to Voicescape, a software which analysed tenancy behaviour, such as the likelihood of residents falling behind on rent. A human decision-maker was in place to make the final decision. She said that she would come back to Councillor Daly on this point.

ACTION: Tina Akpogheneta

Nicola Ellis (Director, Chief Operating Officer, Corporate Services) added that they were looking at using AI to make services more accessible as part of the Resident Experience and Access Programme and would continue to examine this in the second phase of the programme.

Noting that she felt AI could be radically transformational if done right, Councillor Helen Rowbottom asked the following questions:

• Was there any insight about the procurement of AI technology across the sector?
• How could councils capture the value of their data?
• What was the discussion like from a commercial angle?

Sam Nutt replied that the information governance hurdles were significant, and no councils were at the stage of using sensitive data yet. The London Office of Technology and Innovation (LOTI) was working with the Local Government Association, the Equalities and Human Rights Commission, the Information Commissioner’s Office and the Crown Commercial Service on producing guidance on procuring AI in an ethical way. The Equalities and Human Rights Commission’s recent report on the use of AI by local authorities reflected poorly on the meeting of duties, and the barriers in information governance were high.

Councillor Helen Rowbottom followed up by asking how the value of data could be captured both ways, highlighting that councils had a responsibility to capture the full value of technology if it was available. Sam Nutt said that the sector was not quite thinking about that yet, but LOTI was developing data standards and a platform which used AI to find insights from data collected from damp and mould sensors, allowing for more proactive intervention. With regards to Children’s Services, the tools around predictive analytics did not always perform to a satisfactory level, but he agreed that things generally should be moving in the direction proposed.

The Chair asked if the long lead-in time for procurement in the public sector had been considered and if there was any risk of having limited choice if a few big companies were to monopolise the technology. Tina Akpogheneta responded that the use of AI had been embedded in the procurement process and the use of AI by existing providers were being tracked already to balance potential risks and benefits. Sam Nutt added that this was the sensible approach considering the difficulty in making procurement more agile. The Council could still focus on governance and upskilling staff who worked on procurement. Joanna McCormick (Director of Procurement, Commercial and Digital) agreed that the procurement process could be long but pointed out that they had scope to buy digital tools via the existing framework, adding that it was exciting to hear about the guidance LOTI was working on. It was also noted that currently all contract reports went through a lead in Digital to ensure that proper checks had been done before they reached Council members.

Councillor Rory Vaughan wondered if the focus on governance would slow the Council down in the adoption of AI. Referencing a Gartner article, Tina Akpogheneta said that organisations that wanted to be successful with AI needed to get their governance right first to avoid major issues in the future. Generative AI was quite new, so its functions had to be well understood first. Individual licenses that were being tested showed little benefit to the organisation as a whole. The goal was to scale up to tools that could benefit the whole organisation rather than spending money on licenses that could only benefit a few. It was noted that generative AI was not the only solution and there was value in integrating different systems and tools.

Councillor Rory Vaughan followed up by asking if there was an issue with risk appetite with using different types of AI and when the committee could expect to know about the limits of using AI in different services. Tina Akpogheneta replied that the current focus was shaping the approach with SLT, so the team and the leadership could work together to shape the strategy. As an organisation, Hammersmith & Fulham had not always been a first adopter, but the desire was to be able to take advantage of the opportunities offered by AI while managing the risks.

The Chair was keen to know more about the use of AI to assist efforts to tackle fly tipping at Westminster City Council and if there were any lessons that could be applied to the H&F Love Clean Streets app. Sam Nutt replied that the Westminster City Council tool emerged out of a political commitment to reform waste reporting services, and a lot of resources were invested to develop the app in-house. The lesson that other councils could take from this was to have a very clear outcome in mind during the design process.

Councillor Florian Chevoppe-Verdier (Lead Member for European Co-operation and Digital Innovation) addressed the Board and thanked officers and members for their valuable contributions. He noted that the upcoming Data Bill was well-positioned to make the most of AI and he was confident that the Council would be able to balance making savings and improving the delivery of public services while also ensuring residents’ safety. The Digital Accessibility Group (DAG) tested all digital tools being rolled out to ensure accessibility and there were checks and balance to make the most out of technological development. He was confident about the ability of the Council to continue on the AI journey.

Councillor Rowan Ree (Cabinet Member for Finance and Reform) thanked the Board and the officers and guests for an interesting discussion and said that AI could have the same transformational power as computers had in the office environment. It was important to get the basics right at this moment so the Council would become well-placed to take advantage of it. He echoed Councillor Rowbottom that the Council had a responsibility to use technology where it benefitted residents. From a finance perspective, the Council had a responsibility to use technology to make it cheaper to provide services to residents. He agreed with Councillor Daly that the output of AI technology was only as good as the input so using the right data with safeguards was important. As Councillor Perez mentioned, maintaining residents’ trust in the Council’s use of AI was important, following the approach of doing things with residents but not to them. He was grateful for the framework laid down by officers and the expertise of Councillor Chevoppe-Verdier.

The Chair wrapped up by highlighting how far the council had come. Since the last meeting, a proper governance structure, a business case process and a procurement process had been put in place. She said the Board would like the next update to take place once scaled up uses of AI were in place.

ACTION: Tina Akpogheneta

The Chair thanked Sam Nutt for attending and members and officers for their contributions.

RESOLVED

1. The Policy and Oversight Board noted and commented on the report.