The use of AI is spreading rapidly but how can universities keep up?
Right now, universities are being faced with seismic changes due to the pandemic, but another game changer is on the horizon; the rapid rise of ‘industry 4.0’ technologies. The University of Birmingham is adapting to these new technologies. Here's how...
How ‘industry 4.0’ technologies are changing our research
Industry 4.0 technologies offer new opportunities to push back the frontiers of knowledge and uncover rich new insights across all disciplines. I expect that these technologies will be widely embedded across our university in the next 10 years from humanities to engineering.
At the University of Birmingham, adoption of these technologies is still in the early stages, but the use of artificial intelligence (AI) in particular is certainly spreading.
For example, we have groups in linguistics who use AI to process natural language, texts and Twitter messages, and our Health Data Research Hubs are using anonymised large-scale data and advanced analytics – including AI – to develop new insights in disease detection, diagnosis and treatments.
But we’re just getting started. I think the key element of expanding the use of these research 4.0 technologies is going to be how we make the transition from interested groups, such as computer scientists and small teams dotted around the university, to making it more widely available.
This will include creating interdisciplinary links through joint appointments and considering the ethical implications that arise. We will also need to provide continuing professional development (CPD) for researchers across the campus to understand the capabilities of AI in their own fields of research.
We’ll need to pilot AI applications to explore new possibilities and this will require additional computer hardware and the use of robotics to automate some experiments and measurements. I believe that the best way to make a transition in a university environment is for a few people across different research areas to start doing it – these may often be early-career researchers - which will prompt others in their immediate community to turn around and see the potential.
People have been writing books about artificial intelligence for countless years, but it's the hardware capability and knowing how to apply it that is really making a difference now, together with remarkable advances in data handling capability.
Managing research integrity and ethics risks
Research transparency is one of our key concerns.
The increased use of AI prompts us to think about the way that data is used and interpreted and the ethical implications. We are very conscious that a lot of artificial intelligence work is done with a kind of ‘black box’ methodology, where researchers no longer know how the algorithm comes to a particular conclusion. It is a matter of debate as to what extent we should implement these findings in healthcare or other research disciplines.
To that effect we’re currently exploring a much more transparent form of artificial intelligence with a newly-recruited group of researchers who are building new AI approaches incorporating statistical models to get a better understanding of how outcomes are generated.
As member of the Engineering and Physical Sciences Research Council (EPSRC) Strategic Advisory Network, I was asked what the most important topic for the future of research will be, and I suggested that using AI to design improved and efficient experimental methodologies and analysis is going to be vital in many fields to enhance research scope, quality and reproducibility.
Our university needs to invest in these research 4.0 technologies and the capabilities of our researchers to strengthen our international reputation. I don’t think we can take a leading position in research if we don’t invest seriously in these technologies and help our researchers to deploy them across the board.
What technology could mean for the role of the researcher
When automation will be used more widely and replace things people are doing manually at the moment, we clearly need to think about what are the skills and roles that we will need from our technical staff to expand the capability of our research, taking advantage of the research 4.0 technologies.
At Birmingham we have established a Technical Academy, through which we will provide opportunities to technical staff to extend their skills, preparing them for the future research and teaching needs of the University. We will also need to expand our teams of individuals with expertise in software engineering and robotics.
Another element to consider is to make sure academics are on board with the transition and the use of AI and automation. For example, many researchers are fearful of AI taking over peer review, the writing of publications and grant proposals. Of course, this can be seen as reducing the burden on researchers, but there is a very strong feeling within the academic community that it is academics that should be doing these things as there are risks of supressing creative thinking and introducing systematic bias in assessment through automated processes.
That is one aspect where the academic community is a long way off coming to terms with the possibilities offered by Research 4.0.
The impact of automating parts of the research cycle
This is an area where industry and the university can learn from one another. But in certain quarters of industry the use of robotics and AI is more advanced than we routinely use here. We’ll probably benefit from learning what industry is doing and we may need to think about increasing the number of secondments that we have into industry, and increasing mobility of researchers throughout their careers across industry-academia boundaries.
But when it comes to industry partnerships, we have to be aware that the open data (and open software) agenda is not necessarily compatible with the needs of industry to protect intellectual property especially at the more competitive end of research and innovation.
Birmingham's research strategy
The future of research at Birmingham requires us to be clear about and to invest in our research strengths. That way we can facilitate trans-disciplinary collaboration within our university and externally.
We’re aiming to create a culture that attracts the best researchers to Birmingham who can flourish here. We need a backbone of world class infrastructure and professional support, and a culture of individual development and training, as well as one of mutual support between academics.
And as we develop those areas, we also need to embed the concept of responsible and sustainable research and innovation. This is of great importance in contemporary society and applies to universities like ours as well as to industry. I’m conscious that the Research Councils are also thinking about responsible innovation and its importance, in areas from AI and machine learning to synthetic biology.
We need to make sure that these new aspects of research integrity will be part of our strategy and will be part of our training and awareness raising and that we flag them as a consideration for every research project.
What keeps me up at night
Technology developments have been so rapid in the last ten years with the expanded capabilities through data science and AI, that our current pace of change across the sector is not fast enough.
Universities are not always the quickest institutions to react to changes in the world, and our national funding system does not always help with agility. In the past, the longevity of commitment to particular strategies by universities has been a strength. However, in the Research 4.0 context this is also a potential weakness. Keeping pace, particularly with developments in industry, and with the international competition is crucial.
I wonder whether, as a sector, we can invest the resources quickly enough and recruit the right people to actually maintain our competitive position, and be agile enough to move with the times.
Using leadership to improve research performance in a time of research 4.0
We are currently setting up an interdisciplinary Data Sciences Institute which will be launched this calendar year. The idea is to form a centre that brings people together from different disciplines to share expertise, learn from each other and facilitate collaborative work.
Our strategy will be multi-stranded to really take us into a leading place, bringing people together, and taking advantage of our external links with the Turing Institute and its constituent universities, with Industry, and with policy-making organisations locally, nationally and internationally.
For more information about the impact AI is having on the UK’s research sector, read Research 4.0, Research in the Age of Automation. This new report is delivered by independent thinktank Demos and supported by Jisc.
0 Comments