It takes courage, from multiple perspectives
Initiating something new, particularly in the midst of change, at the local, national and global levels, takes courage. I would also argue that truly sustainable change happens when we bring multiple perspectives, disciplines and sectors together around a challenge or opportunity.
That’s why UK Research and Innovation (UKRI) was formed, and Innovate UK is part of it. It invests over £7 billion a year in research and innovation by partnering with academia, industry and government to make the impossible, possible.
UKRI will ensure the UK’s research and innovation system is fit for the future and able to respond to environmental, social and economic change on a global scale by:
- Supporting the development of new ideas and technologies to address the complex challenges facing all societies around the world
- helping the UK to make the most of its world-class research and supporting its businesses to stay at the cutting edge
- promoting public discussion about research and innovation
This brings together researchers and innovators across disciplines and sectors including engineering and physical sciences, arts, humanities and social sciences, the natural environment, biological sciences, among many others.
UKRI has been supporting AI research and innovation for a number of years and we have seen developments in that area increase in number and complexity. We have been looking closely at the opportunities for AI research and innovation in the short- to long-term.
AI opportunity: improving business as usual
I think a lot of AI developments and deployments are targeted at improving specialised functions in business, for example, for optimising manufacturing processes, or personalising retail services. This is all incredibly powerful but I don’t think we are yet seeing adoption of AI at scale in the more general business functions such as HR, legal, contracts, and finance.
A couple of examples of businesses we have supported spring to mind as examples of how AI businesses can grow by meeting these generic needs.
MeVitae – uses NLP to anonymise CVs to remove unconscious bias without losing textual context. This can support equality, diversity and inclusion in recruitment and could have both an economic and social benefit.
More recently, we have started to fund Combine AI to undertake industrial research to develop a prototype AI-based virtual assistant to derive an understanding of the characteristics of virtual meetings and to offer productivity enhancements for meeting stakeholders. I don’t know about you, but I recognise that as a need.
There are major opportunities for greater AI developments and adoption in general business functions that could impact huge swathes of the UK industry. We need to continue AI research and innovation in specialised areas but help businesses to identify and initiate AI development and/or adoption in more general tasks that impact hugely on productivity and resilience.
AI blocker: data isn’t ready
I think many businesses are put off initiating an AI project or programme because their data is not ready. Around 80% of the cost of an AI project can be sunk into getting the data ready for AI, for example, obtaining it, cleaning it, and organising it into training and testing sets.
There is a role for research and innovation in this area to increase productivity in this phase either through cutting costs or accelerating the process and therefore generating value from AI faster.
An AI team needs a diverse set of skills
The last opportunity I wanted to mention is around bringing together a diverse set of skills needed for the successful deployment and adoption of AI. This needs to be done right at the point of initiation but too often it happens too late.
In the case of AI, there are three perspectives I’d like to flag up as particularly important: design, social sciences and responsibility.
AI and a design approach
In design, Innovate UK has long recognised the value of design to businesses, particularly at project initiation. Good design can boost businesses and the economy by enabling the delivery of more valuable outcomes, reducing the risk of innovation, accelerating scale-up and improving business performance. We recently launched a 4-year design in innovation strategy.
To illustrate this point, I’ll mention Qumodo, an AI business which launched in 2016. The team there are dedicated to using the latest technologies to address social issues, particularly in relation to criminal justice and policing.
They recently developed an image search platform for police investigations that is not only easy and intuitive to use, but also reduces psychological distress for officers working on sensitive cases such as child protection.
A human-centred design approach, supported by an Innovate UK grant, was essential to this. The team at Qumodo interviewed a broad range of users in order to create the optimum user experience. They explored and tested alternative solutions before producing a compelling, video-based demo of their proposition.
The company has since won a £1.3 million contract with the Home Office to deliver software helping police officers in the fight against online child sexual abuse and exploitation.
AI and the social sciences
One of the programmes UKRI is running is on Next Generation Services consciously brings AI and social science perspectives together to explore the art-of-the-possible and to ensure AI developments in areas such as legal service, accountancy and insurance take account of the social context and are successful.
One of the projects in that programme is specifically looking at how best to train people in the service sectors to use new technology in their existing jobs and new jobs that emerge over time.
One of the work packages of the project, led by the University of Oxford, is focusing specifically on identifying and testing solutions for gaps in skills and education so that people working in law can make effective use of AI technology.
Responsibility and ethics in AI is a topic that is getting a lot of attention and rightly so. Strongly related to social sciences but I am bringing it out by itself here. This is another area that needs consideration at initiation, even if that means not going ahead with the use of AI when automation is all that is needed.
Responsibility in innovation is not just a matter for AI, of course, and applies to any technology. With that in mind, colleagues at Innovate UK have worked with the British Standards Institute to develop a Publicly Accessible Specification (or PAS) on Responsible Innovation and I would implore you to take a look at it. It is number 440.
There is a lot to think about when initiating a journey with AI. The opportunities are huge and exciting but success will much more likely if a wider set of perspectives is brought in and brought in early.
You can follow Zoe on Twitter
You can go to the Innovate UK website