The potential for using artificial intelligence (AI) to address some of the most serious social challenges of our time seemed clear to the team at Nesta, the innovation foundation. But in order to understand how computer-aided decision-making might be applied they wanted to see how it was already being harnessed in Scotland.

In 2019, Nesta in Scotland set out to find organisations and individuals working with AI to improve social issues such as reducing loneliness or making education more accessible, supporting young people or boosting health.

Nesta’s AI for Good programme attracted applications from more than 70 projects, and £15,000 grants were awarded to seven of them in January 2020. 

“We wanted to understand what was the real lay of the land in Scotland when it comes to AI,” says Adam Lang, head of Nesta in Scotland, which is based at the Bayes Centre in Edinburgh.

“On paper, we have fantastic research institutions in our numerous universities and a supportive government which is developing an AI strategy.

“But we wondered what could be done to harness these technologies for social good and by that we mean delivering benefits to citizens and communities via public services, charities and environmental organisations.

“We, as a team in Scotland, have an interest in digital and technology-driven innovation, so we devised a programme to help us identify, shine a light on and showcase some of those innovators working with AI for social good.”

As a result Nesta found a broad range of projects from across Scotland, covering various sectors, many of them collaborations drawing on academic or commercial expertise in partnership with charities or the public sector.

“We had over 70 applicants which was an incredible testimony to the breadth and depth of interest,” says Lang.

The grant was only part of the plan; the selected seven projects also received tailored support for their work. Had it not been for the huge disruption of Covid, there was also to have been a programme of public engagement from Nesta. 

“We intended to use our convening power to showcase these works, raising awareness of the different ways AI can be used for social good,” says Lang.

Although the plans were scaled back, the project report has been published and it showcases AI’s potential for social good. 

AI is fast evolving; it is already shaping our world with algorithms determining how we are marketed goods and services on Facebook, Amazon and Google, helping fly planes and in time driving our cars, and Lang sees a key role for Nesta. 

“We are keen to champion the role of citizen engagement in the AI process – not just in how AI is delivered, but how it is designed and developed.

“It is crucial that we continue to challenge ourselves and actively engage in that process in Scotland because otherwise you end up being shaped by it, rather than helping to shape it.”

He points to a project at Heriot-Watt University, which focuses on the existing gender stereotypes of virtual assistants. “I think that is why we were really taken by it. There is always a danger that technology like this can be elitist or exclusionary. The Heriot-Watt project stood out because it was about recognising, from the consumer side of AI, that many of these assistants don’t do well in their interactions. 

“We have to ensure that equity, fairness and inclusion are all part of the development of these tools so we don’t you end up with products that are unintentionally discriminatory.” 

Lang sees a positive future. “There is a real potential for using AI tools and AI systems to help us be smarter and more insightful in terms of addressing the challenges we face in Scotland. If we at Nesta can play a small role with this AI for Good programme and showcase some of the potential, then we’ll consider that to be a win.” 

CASE STUDY: BLACKWOOD HOMES AND CARE 

How smart use of energy data will help people with care needs

The simplest ideas are sometimes the most effective. In the case of Blackwood Homes and Care, it was the notion that a person’s electricity use could flag up the need for help. Blackwood cares for and supports disabled people in 1,500 homes across Scotland and responding to calls for help are an integral part of its role.

By taking the data collected by an electricity smart meter and using machine learning to establish what a resident’s normal habits are, unusual activity can be picked up and carers alerted. Called Smile (smart meters for independent living evaluation), the project with Edinburgh University and The Data Lab – the national innovation centre for data and AI – is one of the recipients of a Nesta AI for Good grant. Technology is nothing new for Blackwood, as Anne Jenkins, the company’s innovation delivery manager, explains. “A number of years ago we recognised that technology was becoming more of a factor in determining how you deliver services and improve outcomes.

“We were looking at replacing call systems in our care homes but the options out there didn’t do all the things we wanted to do. So we went down the bespoke route to create CleverCogs.”

CleverCogs allows Blackwood customers to video call, sound an alarm, see their care rota and access the internet. There is an in-house digital skills team that carries out assessments and works with customers to show them how to use the equipment.

“We can see the difference that makes for us as an organisation but also the benefits it gives our customers. Somebody able to open their blinds when they want to and not have to wait for a carer is a really small thing but massive to people living independently.”

Ultimately, Smile can become part of CleverCogs but, at the moment, it is still at the development stage. “We’ve installed equipment with 12 customers and we have another 20 to install – when Covid allows,” says Jenkins. “The project was meant to be running till March 2021, but we’ll have to extend that to get a better picture.

“We haven’t pulled any patterns out of the data; we’re still building up the learning. The algorithms are still being developed. It’s a small pool and unfortunately with Covid we’ve not been able to move that on as much as we’d like.”

Jenkins is, however, excited by the possibilities. “At the moment care is delivered in a reactive way: something happens and somebody calls for help. 

“This gives us the opportunity to prevent some of that. You can see patterns of decline – if someone is taking longer showers or the fridge hasn’t been opened – that might suggest intervention before it becomes really urgent. Smile will give us the opportunity to monitor someone’s well-being in a really unobtrusive way – you don’t have to ask questions, you don’t have to probe.”

AI for Good: the projects

The Citizen Literacy team begin mapping out the learning design for its AI-powered app

ALLI-CHAT An AI-powered chatbot for young people to discuss mental health issues is being developed by a partnership between tech company Voxsio, NHS Forth Valley and groups of young people from Stirlingshire.
CITIZEN LITERACY A group from City of Glasgow College is using AI to develop voice recognition software for regional Scottish accents in order to support adult learners improving their literacy.
DIABETES TREATMENT Red Star AI, a Bishopbriggs technology firm, is using AI analysis of clinical notes and case histories to improve the diagnosis and treatment of diabetes patients.
GREEN MAPS Riverwoods is a joint project between Edinburgh satellite firm Space Intelligence and the Scottish Wildlife Trust which uses AI to interpret satellite data and to map habitats. This will help inform actions to restore and protect Scotland’s natural environment.
INDEPENDENT LIVING Data from smart energy meters is being analysed using AI to monitor changes in patterns of residents’ behaviour. It is being developed by Blackwood Homes and Care, a specialist housing and care provider, working with Edinburgh University and The Data Lab.
MUSCLE POWER Edinburgh University researchers are developing AI-powered prosthetic limbs to help amputees have greater control of everyday tasks, such as gripping a pen or picking up an egg.
SMARTER ASSISTANTS Heriot-Watt University’s Interaction Lab is using AI to address gender stereotypes in conversational assistants, such as Alexa and Siri, by designing and testing new personas and adapting the assistants’ responses.

“Powering Good: Insights from Nesta’s AI for Good programme” is available on the Nesta website www.nesta.org.uk