Artificial intelligence strategy: trustworthy, ethical and inclusive

Scotland and AI – The Case for Action

The why, how, who, where and when

Scotland and AI go back a long way. It's an association that has proved successful across research and business, and one that has the potential to achieve even greater things in the years ahead. That's an exciting proposition and through the Strategy, there is an opportunity to strengthen our AI ecosystem further.

We have our colleges and universities to thank for the thousands of talented individuals now skilled in AI^[9]. The University of Edinburgh is home to the UK's longest-established AI research centre and its School of Informatics is one of the largest in Europe^[10]. Together with Heriot-Watt University, it is a key player in the Data-Driven Innovation Programme^[11].

In the west of Scotland, Glasgow's Turing Institute was an early pioneer of AI development in the 1980s^[12]. The baton has now passed to the University of Glasgow, which plays a key role in iCAIRD – the Industrial Centre for AI Research in Digital Diagnostics, a collaboration of 15 partners across Scotland including the universities of Aberdeen and St Andrews^[13], the NHS and industry.

Scotland is also home to several hundred businesses, from start-ups and SMEs to major corporations, actively developing, adopting and applying AI to identify new drugs^[14], diagnose disease^[15] and improve animal health^[16]. These companies have a global voice – the world listens to what we have to say about AI. Opportunities to collaborate with our international partners will only increase as our reputation grows.

Today, we have ambitions to use AI for positive effect across all society. The public sector is adopting AI to create better economic, social and environmental policy and interventions, and to provide more intelligent and targeted responses.

In Scotland

7 out of 16 universities offer AI degrees at undergraduate and postgraduate levels

2 offer it at undergraduate level (Edinburgh and Robert Gordon University)

7 offer it at postgraduate level

960 students are studying it at undergraduate and postgraduate levels

565 students study it at undergraduate level

395 students study it at postgraduate level

Based on 2019-2020 admissions

In Scotland, there are 2 dedicated AI Centres for Doctoral Training (CDTs) and an additional 3 that include elements of AI.

We are healthy and active

Industrial Centre for Artificial Intelligence Research in Digital Diagnostics (iCAIRD)

The Industrial Centre for Artificial Intelligence Research in Digital Diagnostics (iCAIRD) is a pan-Scotland collaboration of 15 partners from across industry, NHS and academia. With funding from Innovate UK and key industrial partners, iCAIRD is building a world-class centre of excellence focusing on the application of AI to the field of digital diagnostics. iCAIRD enables research-active clinicians and innovative SMEs to collaborate on key clinical questions and ultimately solve health and social care challenges more quickly and efficiently.

iCAIRD is initially focused on building infrastructure to host an AI development environment for radiology and pathology, starting with research hubs in Glasgow and Aberdeen, as part of a national platform for federated learning. This covers skills development in data and security governance processes and data extraction from local and national NHS archives. It also creates platforms to facilitate AI training and experimentation and provides support for clinical evaluation and validation.

An important element of the infrastructure is the involvement of exemplar partners to help build AI capability in areas including radiology imaging (e.g. acute stroke, chest x-ray and mammogram) and digital pathology (e.g. gynaecological cancer screening). As the programme enters the sustainability phase, iCAIRD is examining ways of supporting new projects and model generalisation through its network of federated training sites. A key aspect of this work is patient and public engagement to address issues of trust and ethics which surround AI's place in healthcare.

icaird.com

While Scotland can look to the future with confidence, we recognise that AI brings with it serious challenges. For AI to be truly inclusive and benefit everyone, we need to secure people's trust in the technology and be clear on its role in our society. We need to address the real risks and concerns of bias arising from inadequate data or design or a lack of transparency of decision-making. This is an opportunity for Scotland to lead the world in ethical AI, building on work already happening. The Edinburgh Futures Institute^[17] is focused on helping society navigate complex futures, and is investigating the ethical implications of advances in AI in applications spanning agriculture, education and healthcare. It is developing frameworks for building trustworthy systems, and offers an online course on Data Ethics, AI and Responsible Innovation^[18].

As we come to terms with the socio-economic issues brought about by COVID-19^[19] and the adoption of AI in our lives becomes more prevalent, our Strategy is designed to make it a force for good and a catalyst for positive change.

Our Digital Strategy sets out a path for us to address together the unique challenges that AI creates, and ensure Scotland fulfils its potential in a digital world.

We are well educated, skilled and able to contribute to society

AI in the Public Sector - Data Science Accelerator Programme

The Data Science Accelerator Programme, part of the broader Scottish Public Sector Analytical Collaborative (SPACe) programme, is playing an important role in improving vital data science skills in Scotland's public sector. It allows analysts to develop their expertise in this area, under the guidance of an experienced mentor, using AI techniques such as machine learning and natural language processing.

The Scottish Government launched the scheme in 2018 in partnership with Public Health Scotland (then NHS National Services Scotland), the National Records of Scotland, Registers of Scotland and The Data Lab. It has since reached out to many more public sector organisations, such as Police Scotland and the Care Inspectorate, to make the programme available to all organisations across the country that produce and publish official statistics.

The programme has successfully taken forward a diverse range of public sector projects such as the use of machine learning techniques to determine whether satellite information could be used to improve agricultural statistics. Another focused on the extraction of information from historical images using computer vision and machine learning techniques to generate valuable information on the built environment at that time.

The Care Inspectorate, in an innovative proof-of-concept project, explored the use of algorithms to identify poor quality care at an early stage, information that could then be flagged and passed to inspectors for further assessment and action.

And in a recent project, participants applied natural language processing techniques to develop an opinion tracking tool to monitor comments on social media relating to social security benefits in Scotland. A successful proof-of-concept has led to an extension of the project beyond the Accelerator Programme

A new land cover map of Scotland for 2020 produced through combining multiple different satellite datasets using Space Intelligence's advanced AI system

We value, enjoy, protect and enhance our environment

AI and Satellite imagery to tackle climate challenges

Space Intelligence

It's said that climate change is the main challenge of our lifetime. There's a greater need than ever to accelerate our zero-carbon transition and protect the natural environment before it's too late. AI is playing an important role in this race against time, providing scientists, governments and conservationists with the data they need to make informed decisions and develop progressive, net-zero strategies.

Significant biodiversity loss is occurring a result of human activity, specifically changes in land use. However, the emergence of Nature-Based Solutions (NBS) allows the environment to fight back by conserving and replanting forests and restoring our peatlands. However as the majority of NBS sites are large and often remote, the only cost-effective way to monitor them is by using satellites.

Space Intelligence is a specialist in satellite data analysis and undertakes projects around the world, some of them close to home in Scotland. Supported by the Scottish Government through the Can Do Innovation Challenge Fund, Space Intelligence developed an AI system that runs on its satellite analysis platform to map Scotland's land cover on an annual basis. Partnering with NatureScot, the system identifies and tracks changes to the ecosystem year-on-year, providing stakeholders with evidence of environmental deterioration or recovery – information that confirms if Scotland is on track to meet its climate change targets. In a similar project, Space Intelligence was funded by NESTA to use AI to map riverine forests with the Scottish Wildlife Trust, to support planning a Nature Recovery Network. The company has also recently been awarded a place on the Data Driven Innovation (DDI) Post COVID AI Accelerator Programme.

https://space-intelligence.com/