AI

AI – progress to a local government future

Following a discussion last year on what AI promised for the councils of the future, The MJ and Penna reconvened a group of experts in the field to discuss progress so far. Paul Marinko reports

© Phonlamai Photo/Shutterstock.com

© Phonlamai Photo/Shutterstock.com

Time waits for no one – just ask Rishi Sunak… And with that in mind, The MJ and Penna pulled together a selection of sector figures to chew over how the sector was faring in its pursuit of embracing AI.

There was general agreement that while a tiny few had dived in, most had just begun to tip their toes into the water.

‘People are beginning to understand what it means,' said one participant.

While the benefits of a little bit of using Microsoft's Copilot was beginning to be experienced by some councils – such as summaries of meetings and help with form-filling – many were still shying away.

‘Most councils have not deployed it because they haven't felt there is a strong enough business case,' said another, who felt the cost of licences was creating a nervousness around adopting it.

And while many services such as social care were calling for solutions, there were emerging capacity issues within internal IT teams that risked creating bottlenecks in implementation.

Another risk identified by a participant was that of councils seeing AI as simply a tool, when in fact it promised ‘a new way of service delivery'.

They said councils should be focused on what their service delivery model was and what solutions they were looking for.

But, while it could be easy for councils to get carried away with the transformational potential of AI, there was – according to another participant – the need to make sure residents had trust and confidence in any deployment of AI by their local authorities.

While most councils were transfixed by the thought of doing something with Copilot, they were missing that this was just one example of AI from which they could be reaping benefits.

In some countries, they added, there would be 200-plus AI tools deployed the second you stepped into an airport.

‘It's not just about Copilot,' they concluded.

This was supported by a fellow guest, who emphasised there was often a great deal of AI capability within the software councils was already using that were not being utilised.

‘Councils need to make sure it's turned on because there are huge opportunities to get benefits without the need to work up a full business case [to utilise new AI tools],' they emphasised.

Another agreed there was ‘still a long way to go' and local authorities needed to stop seeing AI as a ‘new shiny toy' and shift to a ‘outcomes-focused mindset'.

But, while accepting this point, one voice around the table emphasised the need to recognise the ‘merits in taking baby steps' with this level of change in councils. Showing the simple transactional benefits initially would allow the organisation to slowly recognise the wider strategic potential.

This suggestion of baby steps was enforced by another participant, who argued there were three phases for any organisation adopting AI. The first was ‘foundational', the second was ‘transitional' and finally there was ‘transformational'. The truth was, they argued, councils were only at the ‘foundational' stage.

There was a need for them to ‘clarify the ask' so the organisation was ‘really clear' about what it was trying to achieve from adopting AI. And ‘technologists' in the organisation needed to be leading this.

But, while it could be easy for councils to get carried away with the transformational potential of AI, there was – according to another participant – the need to make sure residents had trust and confidence in any deployment of AI by their local authorities.

With councils already facing a public backlash over a perceived growing ‘facelessness', this was not a risk to be ignored lightly.

Yet there was broad agreement councils could be pretty good at finding excuses for why it was difficult to fully embrace AI. And data concerns were a prime culprit for this.

As one participant put it: ‘There's a tendency to let the barriers get in the way of the solutions.'

Cyber security was roundly recognised as a risk and there was general agreement that breaches would inevitably happen from time to time.

‘The issue with AI is that chances are there will be some data breaches,' said one officer. ‘But we are able to mitigate this factor.'

‘We can't eliminate all risk,' agreed another. ‘We just need to ensure we are ready for it, because AI is going to be used by bad actors. It's an increasing risk.

‘We need to be partnering with people who are experts in this. And end-users are the greatest risk.'

But partnering and collaboration was clearly a sticking point, with major disconnects occurring even within individual organisations.

A fellow participant concurred, adding: ‘There is a massive gap on data standards across the sector. We need a common set of standards.

Skills was also raised as a hinderance to AI being effectively implemented in councils, partly in the face of pay challenges. And this had the potential to clash with expectations – especially among some local politicians – to move at speed on implementation.

But one voice around that table added: ‘It goes further than skills. There's also a cultural piece.

‘We need other services to be thinking about this, but they are looking to the corporate centre to provide it.

‘We need to bring in HR and OD to manage that.'

Externally, there was evidence of a clear desire to head in a collaborative direction, with round-tablers believing there was a ‘genuine desire' to collaborate and some thinking the sector was ‘just on the cusp' of a new dawn in this regard.

‘If there's a desire to come together that's a good thing,' said one. ‘The questions is: Can we find a few things we can work together on?'

The point was also made that resources were making the need for collaboration a ‘reality'.

But there was a call by another participant for a different reality to be recognised. They felt it was too hard for councils to collaborate with each other at the moment.

‘It needs to be government-led because we are so siloed,' they argued.

‘We have to face facts. There is limited time, space and resources. It isn't real at the moment.

‘Shared capability is what government can be bringing to the table.'

They added that councils were all different organisations and it wasn't possible to ‘align a solution for all of them'.

‘By all means plagiarise the life out of all this, but it's not going to work as a central procurement function.'

Many concurred, with one pointing out that the only cross-council collaborations that worked were ones where a single council took the lead and the others followed on with little interest.

So, to summarise, AI is definitely here, but it certainly hasn't built up a head of steam yet.

With everything from capacity issues, lack of standardisation, skills gaps and cultural issues to misunderstandings over AI's potential and fears about the risk of embracing the technology, there is plenty holding back progress.

But, AI is not only here, it is also here to stay. So, councils are going to have to up their game at some point. w

 

Comment by Bruna Varante, consultant, executive search team at Penna

Earlier this year, I had my first experience with AI in local government. Yes, Chat GPT and some other tools have been around for a while now, but I have always been a little sceptical about using it. However, one day I joined a call to discuss one of my projects and while I was waiting for consultants to join, ‘Otter' joined without me or my colleague, having authorised it. Neither of us knew who ‘Otter' was.

We soon realised that one of our consultants had an Otter.ai note-taking tool to help with the minutes and actions. Surprisingly,it worked really well. It summarised the key points of the meeting, differentiated ‘soft' chat from the essential actions, and gave us all a really compelling piece to refer to in our next meetings.

I was pleased but also intrigued by what that tool meant. My negative bias towards the unknown was evident. I could realise the short-term benefit. But, could that mean something else in the long run?

Maybe just a note-taking tool wouldn't harm it. But what if the note-taking could be linked to a broader network of AI tools, where the data is interconnected and potentially vulnerable to breaches or misuse? Could that be the beginning of the dangers we cannot yet clearly see?

Councils deal on a day-to-day basis with an infinite amount of personal information, and we can see from the recent NHS cyber attack that no matter how safe organisations are, we are still far from saying they can be 100% secure.

In a world where data is its most valuable asset, we must walk cautiously. AI can have a huge benefit in improving the quality of services provided to residents. But to do that, we need to be as quick in providing the tools, strategies and frameworks required for its safe implementation.

 

Round table participants

Mark Lumley – director for digital, IT and resilience, Hounslow LBC

Rehana Ramesh – acting director for customer service, digital and data, Hackney LBC

Ian Robinson – chief digital and information officer, Richmond and Wandsworth LBCs

Omid Shiraji – chief information officer and adviser, East of England Ambulance Service NHS Trust

Gavin Beckett – chief product officer, Placecube

Nadira Hussain – chief executive, SOCITM

James Johns – head of UK corporate affairs, Workday

Amin Aziz – senior consultant, Penna

Rachael Morris – senior consultant, Penna

Matt Jones – senior consultant, Penna

Zeynep Livatyali Esen – senior consultant, Penna

Tristan Moakes – strategy director, Penna

Bruna Varante – consultant, Penna

Heather Jameson – editor, The MJ (chair)

Paul Marinko – deputy editor, The MJ (reporting)

AI

Curing the hybrid headache

By Matt Etherington | 11 October 2024

Matt Etherington looks at how local authorities have been tackling hybrid workspace challenges

AI

Joining up relational regulation

By Jess McGregor | 11 October 2024

The early days of a new government offer a chance to consider what collaborative and purposeful assurance would look like, says Jess McGregor

AI

Breaking free from short-term fixes

By By Owen Mapley | 11 October 2024

The sector needs immediate support to 'start planning for an alternative to episodic crisis management', says Owen Mapley.

AI

Ensuring citizens' assemblies land

By Graham Smith | 11 October 2024

Graham Smith argues that failing to make citizens’ assemblies effective can do serious damage to democracy and that authorities must be responsive to local v...

Paul Marinko

Popular articles by Paul Marinko