First steps towards effective use of AI in education
Institutions should take time to decide on the best way to make artificial intelligence (AI) work for them. Jisc’s AI maturity model can help.
AI is already showing promise for education in two distinct areas: personalising the learning experience for students and freeing up time for staff by taking on some of the more repetitive tasks.
However, building it into an institution’s digital strategy and teaching methodologies can be challenging. As well as its obvious benefits, AI does present risks and governance issues, and we need to develop new approaches and processes that enable the sector to really take advantage of the things AI can do well.
Implementing AI is not something to rush into. To be effective, it needs careful planning, trialling and rollout. There’s a lot to think about.
Not using AI yet? You’re not alone
Currently, only a small group of universities and colleges have institution-wide AI processes or systems running. Some are just starting to experiment with it, but the majority are still hesitant.
To help the tertiary education sector along the journey to AI adoption, Jisc has defined a model which institutions can use to conduct a scoping exercise.
The AI maturity model makes it easier for institutions to understand where they are, where they want to get to, and what sort of activities might be needed in order to progress towards effective AI implementation.
Approaching and understanding
At this stage, educators will have heard of AI and how it might be used to improve teaching and learning. They’re interested, but not sure where to start. The first step is to understand how AI works, what kind of problems it can solve, and how it is already being used in the education sector.
It’s also important to examine and understand the broader issues around AI. What are the key ethical issues and what wider societal impacts should be considered? For example, how might AI’s energy usage affect the environment?
To make sure AI-driven resources are accessible to all, institutions should ask what specific accessibility issues might arise during the use of AI and what opportunities it provides.
Seeking the views of various stakeholders is also important at this stage. What are students’ attitudes to AI? Are they already using it? Do they have concerns? To make sure AI-driven resources are accessible to all, institutions should ask what specific accessibility issues might arise during the use of AI and what opportunities it provides. A useful starting point is Jisc’s AI in tertiary education report.
Experimenting and exploring
The best way to proceed to the next stage is to identify a particular problem to solve, and then explore whether an AI-based solution has the potential to help. We often see people looking at this in reverse: we have all this data - maybe AI can do something useful with it? This approach is rarely successful, particularly in the early stages of exploring AI.
On the other hand, enthusiasm for AI shouldn’t overshadow the true requirements – it may be that a process review is all that’s needed rather than a new piece of software. One thing is certain: data is a key requirement for successful implementation of AI, and it’s important to understand that data maturity and AI maturity go hand-in-hand.
For example, is the data that’s needed for the application available? What data concepts are important to understand and consider? How does bias occur and how can it be mitigated? And can the method that the AI system uses to reach its decisions be explained in a way that users can understand?
Running small-scale pilot projects provides the opportunity to explore AI, establish the right ethics processes and embed a more data-driven culture into the organisation.
At this point it might be worth engaging with us to find out more about Jisc-supported pilots to test AI applications such as chatbots, assessment tools and personalised learning. Running small-scale pilot projects provides the opportunity to explore AI, establish the right ethics processes and embed a more data-driven culture into the organisation.
Operational
Once the experimentation stage has been successfully completed, AI systems can become fully operational for one or more processes institution-wide.
This will require a new approach to understanding and validating the performance of the software. These issues should be fully assessable during the procurement process and then capable of being monitored over the lifetime of the system. The results should be used to inform future developments and requirement specifications.
Embedded
During the early part of the operational stage, AI projects are likely to be seen as special cases but, as AI is embedded, it becomes an integral part of an organisation’s strategy and a serious consideration for any digital transformation project.
Policies for automatic monitoring of the effectiveness of AI models and systems should now be implemented at an institutional level, in much the same way as institution-wide cyber-security policies. And, with mature data governance in place, you’ll see AI start to make a positive impact.
Transformative
Now the right foundations have been laid, AI can deliver on its promise to transform the teaching and learning experience by alleviating the burden of administrative tasks for staff and providing personalised learning for students. And institutions will be ready to take advantage of new AI applications as they emerge – which they will undoubtedly do.
Find out more
- Our work in AI includes running a series of webinars to help institutions progress towards AI maturity.
- Read a blog about Jisc’s AI maturity model.
About the author
I lead our work supporting the responsible and effective adoption of artificial intelligence across the education sector, through a range of pilots, advice, guidance, and community support activities.