The word 'algorithm' is no longer just understood by data geeks; all around the world, people's awareness of data is growing - as is the volume of information being collected about organisations, communities and individuals.
We are increasingly occupied by the many benefits and possible pitfalls of data analysis and artificial intelligence, yet there is a disconnect. While we recognise the human impact, to many, this is a field that sounds clinical and removed from our day-to-day lives.
In fact, it’s anything but. Today, I am Jisc’s executive director of data analytics, but I previously worked as a psychologist - and it’s been a relatively smooth progression from one field to the next, with significant cross-over of skills. Let’s bring these worlds together and look at how mindset can inhibit or help us embrace the potential of digital interactions.
The National Data Strategy (NDS), which published in September 2020, describes itself as ‘a central part of the government’s wider ambition for a thriving, fast-growing digital sector in the UK, underpinned by public trust.’ Trust is a very human concept - and it’s a word that appears in the NDS document no fewer than 51 times, with public trust recognised as a limiter in what technology can deliver.
Perhaps it’s unsurprising, therefore, that when public sector reactions to the strategy were discussed at a recent online event, they called for a holistic approach to the implementation of digital and data skills, breaking down silos, and seeing these skills as intrinsic to the success of organisations’ activity.
For these reasons and more, the NDS marks a positive step forward - but I believe it could go further.
Actively endorsing a set of common data standards – such as those defined by the Higher Education Statistics Agency (Hesa) and Jisc’s own national learning analytics service - would make it easier to trust data and use it to compare performance. This could also free up resource to focus on more productive outcomes, including technical resource to focus on innovation - and allows data-users to focus on what to do, rather than worry over data accuracy. Specifically, the opportunity presented by a student identifier that follows individuals across their learning journey is one we are keen to progress. This has been in the ‘too difficult’ box to date, but that could well change with the Skills Bill’s lifelong learning promise.
A set of common data standards would also serve to reduce the sometimes significant costs of integrating data between sources. Charles Baird, data architect at the Data Standards Authority, has explained why this is important, noting: “most organisations have issues with data management […] because there have historically been no external bodies to set standards for data use.” Emails to clients, databases of plain-text files, and other banks of data are currently implemented in different ways, potentially even within one company - and this lack of standardisation prevents information being shareable and inter-operable. That’s important for universities conducting data-informed research, for example.
There’s also a significant gain to be made by increasing digital and data awareness in education settings. In Jisc’s latest Digital Experience Insights (DEI) surveys, only around half of learners in further education (FE; 54%) and just 42% in HE agreed they had received guidance about the digital skills needed for their course. And it doesn’t stop at the college or university – we need to do more to support learners to transition into Industry 4.0 employment. Research commissioned by the Department for Digital, Culture, Media and Sport found almost half of businesses (48%) are recruiting for roles that require data skills, and that 46% have struggled to recruit for these roles over the last two years.
The DEI surveys also found that only 37% of FE students and 36% of HE students believed their organisation told them how their data was used, despite more than a third of them working with data on a weekly or more basis (FE: 35%, HE: 37%). Jisc’s response to the NDS therefore states: ‘It is vital that any long-term strategy for data and the education sector must include plans to increase understanding and trust in how personal data is being accessed, and ensure learners are being provided with the skills to use data confidently’. This extends to staff too, as the opportunities presented by analytics to support teaching and student support will be limited by their data confidence and digital skills. Trust can be built through having transparent and communicated codes of practice – such as the Jisc wellbeing and learning analytics codes of practice.
Access for all
Digital and data poverty is another area to address with urgency. As teaching, learning and assessments shifted online and off-campus in response to the pandemic, the scale of this issue was exposed – and while universities, government and sector bodies were quick to recognise the need to support students with access to devices, it took longer to see the wider issues and fuller impact. Beyond the need for laptops and tablets, many learners struggle to meet the data costs required to complete their studies, a great number battle with housemates for bandwidth, and – particularly in remote areas – connectivity has been an issue. The impact of digital poverty on equitability of skills and capability remains significant.
The NDS could go further here, supporting creative solutions to widening participation. For example, Jisc is currently working with local authorities, with the support of government departments and the universities minister, Michelle Donelan, to enable free wifi in public places through our eduroam provision. This would deliver much-needed connectivity, without the need to be located on-campus.
'Strip away bias'
Artificial intelligence (AI) can help too. Much of my work as a psychologist focused on working with children and young people who hadn’t had the best life opportunities. In some cases, this was exacerbated by bias - often unconscious. AI and algorithms have the potential to strip away bias from selection and assessment, and can nudge young people to career and life choices they possibly wouldn’t otherwise have taken.
Both as part of my role at Jisc, and as a member of the education advisory panel at the Centre for Data Ethics and Innovation (CDEI), I am focused on the potential opportunities and issues around the adoption of data-driven technologies and AI. These are key questions for Jisc’s National Centre for AI in tertiary education, for the CDEI’s AI Barometer research report, and for its Education Advisory Panel - all of which are collaborative initiatives that seek to drive positive change, uniting industry, government, academia and civil society, while supporting students and staff. They recognise the human interests at the heart of conversations about data - and, for me, that’s the bottom line; we will only achieve our data and digital ambitions if we focus as much effort on people’s interactions as we do on technology.
Perhaps now is the time to look at the emotional impacts and societal benefits, particularly in terms of equality for students of all backgrounds. Let’s build public trust in algorithms and tech. And let’s give people reason to trust good data and good AI.
Find out more about our new centre to lead the way for artificial intelligence (AI) in education.