Roundtable: AI could be an ‘absolute game changer’ for Local Government

Local Government’s AI Journey

“We’re just at the beginning of our AI journey,” says the interim assistant director – customers, policy and improvement at Merton LBC. “But we’re incredibly interested in how we can use AI – in an ethical and sustainable way – to improve the quality of experience for our residents, our communities, and in a way that makes day to day working life easier for our workforce.”

Mr Burns was speaking at a recent LGC roundtable, held in association with Hitachi Solutions Europe, which brought together a small panel of experts in the striking surroundings of the City of London’s second highest skyscraper, Heron Tower. Together, they considered how councils might most helpfully and appropriately harness the power of artificial intelligence. A separate panel considered the same issue during a companion roundtable held in Manchester.

AI, in which computers perform tasks commonly associated with human intelligence, has been making headlines of late. That focus became particularly intense following the public launch of Open AI’s ChatGPT in November 2022. Within two months, the AI chatbot had reached 100 million users – believed to constitute the fastest ever growth in a consumer internet app – and had beguiled with its ability to create human-sounding responses to all manner of questions.

Suddenly, generative AI – models which can create text, images or other media in response to human prompts – had become accessible to millions of people.

The result has been headlines which tend to be at two extremes, either contending AI will transform our lives in all manner of positive ways, or suggesting that without due care and attention we will all be ruled by machines in the sort of dystopian nightmare previously confined to science fiction.

Balancing AI Technology

As so often, the likelihood is that the truth lies somewhere in the middle. “It’s a balance between the ambition of doing a phenomenal number of things differently, and checks and balances for doing it the right way,” said Sharna Quirke, strategic director for local government and health at Hitachi Solutions Europe. “It’s making sure guardrails are in place – that everyone understands where the risks are, but also how to mitigate them.”

And while AI may feel like a bewildering new world, participants were clear that in many ways it is nothing local authorities have not encountered before.

“This is an opportunity for us to do better through using technology,” said Tim Spires, director of IT, digital, and transformation at Oxfordshire CC, who emphasised the need for councils to collaborate in building an understanding of AI. “For me, AI is an extension of what we are already doing. We already have automation, digital, technology, and innovation. So what is AI? It’s another segment of this field. I can’t think of a better way to describe it. I think it’s important that we’re not dazzled by the new toys that are being paraded all the time.”

“AI is the latest, shiniest tool. But I don’t think you should look at it in isolation. I think you should look at how you leverage AI with other technologies to really transform the line of business services.”

That is a process in which staff at Buckinghamshire Council are currently engaged. Sarah Murphy-Brookman, the authority’s corporate director of resources, can envisage all manner of applications for AI.

“I’m particularly interested in its generative and predictive capabilities,” she said. Could AI help public health teams by predicting which citizens are likely to become ill, she wonders? Could it guide social care teams on when care packages will need to be put in place for specific individuals? Could it help identify why certain bins are always missed on refuse routes, and suggest how to redesign street configurations to address the problem?

“For me, there is a lot of AI predictive work that will be an absolute game changer.”

Starting simple

That said, Ms Murphy-Brookman did not feel these are necessarily the best places to start for an authority keen to use AI. “Although I have an ambition of where I would like us to be, it’s also thinking about the low hanging fruit [for AI] that we understand right here, right now.

“So it is complaint responses. It is the driving down the number of calls, it’s freedom of information requests, it’s the production of minutes and actions automatically. It’s that easy stuff where you can test it and model it; you understand the outcomes very quickly and people can become secure, and relatively knowledgeable about this early stage of AI. And then we all go in lockstep together as an organisation about the opportunities it presents.

“If we can get it rolled out as widely as we can starting with those easy things, those low hanging fruit that take out volume and will probably drive cashable efficiencies as well, we can then move into the more bespoke activities.”

Mr Burns suggested another advantage in starting with these simpler areas: it offers more time to explore some of the deeper issues in using AI.

“I’m really interested in generating complaints responses, where you get a large number of complaints that are not dissimilar in nature. I think there are opportunities to use generative AI to do that without having to get too far into some of the ethical issues we’ll need to face. So I think this is where we probably want to start because it feels less complex.”

Data ethics

At Brent LBC, that awareness of potential ethical issues has already led to the creation of an independent data ethics advisory board. It is chaired by one of the council’s corporate directors, but the rest of the membership is mostly external.

“We’ve got an information governance specialist, and a lawyer who’s got a strong data governance background; we’ve got someone from the Greater London Authority, we’ve got someone from the NHS, we’ve got an academic,” said Sadie East, the council’s director of transformation. “So it’s a good sounding-board and reference group.”

The advisory board uses the Open Data Institute’s ‘data ethics canvas’ tool to identify and manage any ethical issues that emerge with data projects. “It gives you a framework to look at what you’re doing and identify whether there are byproduct issues or ethical issues around what you’re thinking of doing.”

The board and council have already encountered such challenges. “We’ve been doing some work with damp and mould sensors in some of our housing,” said Ms East. “That gives us some really interesting and useful information around how people’s boilers are working and where there might be risks of damp and mould. But actually it also shows us where people aren’t turning on their heating at all, which could be an indicator of cost of living problems.”

It provides an example of a key ethical concern with AI: what happens when a council wants to apply AI to data that was collected for an entirely different purpose?

“Collecting new data [to use with AI] is almost the easier bit,” said Ms Murphy-Brookman. “It’s how we handle applying AI over old datasets that were maybe collected years ago, maybe even before privacy notices existed.”

Helping residents keep up

It brought the discussion to a central point: the idea that, if councils are to use AI effectively and appropriately, engagement and education will be vital. In Newham LBC, that has involved setting up a ‘data campus’ in the town hall.

“One of our biggest concerns is around the ethics and regulation of use of AI and public data,” said Nathan Nagaiah, data economy senior lead at Newham.

“We want to make sure we use data ­ethically and effectively but, at the same time, we want to see how advancement in AI and automation can provide opportunity for local residents in terms of future jobs.

“We’ve got a lot of jobs locally that will be affected by automation – lots of local residents work in manufacturing and retail. So at the same time as the council moves towards automation, we want to make sure our residents can keep up with the movement of industry [towards AI].”

The data campus – the first in London – is part of that work. Focused on AI and cybersecurity, it is intended to help young people develop skills and careers. It’s also designed to serve as an incubator for start-ups and businesses in high growth industries.

Positive disruptors

Citizens aren’t the only group which will need to be supported through potential AI transformation, panellists stressed. Within councils, both officers and members will need to develop understandings of what AI means and its appropriate use.

In some instances, that will involve educating staff on something they are already using. “We know we’ve got several hundred colleagues across the organisation who are using ChatGPT for whatever reason,” said Mr Spires. “We dare our staff to do things differently, so we can’t then berate them for trying to come up with something which they think is innovation and achieving better outcomes.”

“The whole question of implementing and delivery gets you back into the world of how we do change in local government,” said Mr Burns. “Essentially, these are change projects, so how do we make sure the organisation understands what’s happening? How do we make sure that if there are disruptors in the organisation – I use ‘disruptors’ in a positive way – that they’re part of the process and not operating outside it in a way that could become disruptive in a negative way.”

At Buckinghamshire, there has already been a briefing session about AI for all those at head of service level and above. “We’ve talked about ChatGPT and the ­possible dangers of putting data out there, and actually we’ve closed down use of it,” said Ms Murphy-Brookman. “We’ve told them they can only use Bing Chat ­[Microsoft’s generative AI chatbot] because that keeps you within the safety of your perimeter wall.

“We have to remember our change management processes,” she said. “We need to apply the same rules [to AI as to any other change]. So we need to go through the initiation stage, we need to go through the discovery phase, then we need to go to the planning and create the ­business case, then we need to get that signed off. And then we implement and we review.

“We don’t just jump in without doing all of that. So it’s still doing all of the normal stuff you would ordinarily do, but in a very exciting space.”

For Ms Quirke “it all comes down to skills”.

“Of course no organisation is going to implement AI and then walk off,” she said. “AI isn’t just implemented. It’s designed into an organisation.”

And that design is, crucially, by humans. “That AI complements humans, that it complements their work but it doesn’t replace, is a really important message.”

“There’s so much opportunity here,” said Mr Spires. “We need to be mindful of the risks and manage those, but harnessing the power of this is an opportunity at a time when we’re in a really difficult place both nationally and locally.”