Our research aims
Conversations around Generative Artificial Intelligence (GAI) and its impact on education continue to dominate conference agendas and fuel fierce debate. As a research team in Group CTO, we are working to understand the daily reality of how GAI tools are being approached and used within institutions and to learn about their impact.
In this blog we share key insights from our analysis of responses to a free-text survey about GAI use. Replies to this survey were invited from learning technologists, library staff and academic developers from across the sector. Survey participants shared their thoughts about GAI, and the level of their experiences using tools – discussing purposes for doing so, the impact the tools are having on their roles and how they anticipate use of them in the future.
Because sector discourse has been so pronounced since the launch of ChatGPT catapulted the subject into the mainstream, it is difficult to find much to say that is novel. The purpose of this blog is not to do that – indeed, many of the insights gleaned from this survey may feel very familiar. Nor is to give advice – the Jisc Artificial Intelligence team provide that service.
Rather, our aim is to facilitate a comparison of notes about GAI usage across the sector. We hope that sharing insights into everyday experiences with GAI will help to ground and inform the conversation and help educators and institutions as they continue to develop their own approaches to GAI use.
Insights – a topline view
Given the strength and extremes of opinion often aired on the topic of GAI and its role in education, it could be assumed that responses to this survey would be variable and vehement. But this is largely not the case. While a few participants who responded to this survey did take the opportunity to put forward strong views (some intensely positive, some fiercely negative) there is otherwise a fair amount of consensus and compromise.
Attitudes towards GAI tools and approaches to their implementation are largely aligned across responses. Similarly, the challenges faced, and concerns raised often correspond; reasons for using the tools and the benefits of doing so are common; and there are widely shared assumptions about the future.
In the rest of this blog, we will unpack areas of difference and agreement in more detail.
The state of play – current attitudes and institutional approaches
There is rich discussion from survey participants about current approaches to GAI in their institutions as the tools continue to evolve and become more integrated with teaching, learning and assessment. While individual attitudes vary – ranging from enthusiastic adoption to cautious implementation – there are clear similarities in the way the survey participants and their institutions are tackling GAI, and the challenges they are currently facing because of it.
A key challenge area discussed in responses is that of inadequate policy for staff and students around GAI usage. For a few survey participants, a complete lack of any guidance is causing concern. But the more common problem highlighted is that policy wording is present but not sufficiently formed or in-depth enough to meet the requirements of staff and students from all areas of the institution. Some of the institutions discussed have published statements around using GAI in an ethical manner, but many responders feel more detailed and considered policy is required for those seeking clarity and guidance.
Library staff highlight that their services and perspectives should be considered when developing policy and guidance. They also view their existing literacy skills as crucial to aid students to navigate GAI usage and strengthen their AI literacy. Some of these staff believe that building their own deeper understanding of GAI will help them improve in doing so.
Learning technologists report being particularly in service to academic staff, providing guidance about working with GAI tools. Participants in all three job roles stress the importance of training and express a sense of responsibility to help students and their peers understand the tools and promote AI literacy and ethical usage. It is important to note that many are willing to take on this role, even if they do not personally feel positive about GAI.
Despite a mixture of positive, negative and neutral stances, there is an acceptance that the integration of some GAI tools will be unavoidable and that ignoring them isn’t an option – simply put, there is no choice but to adapt. Instead of avoiding GAI tools, looking at their affordances and critiquing their output is the approach best favoured.
A small but genuinely expressed set of concerns put forward that there is a need for a contemplative pause to happen now – an evaluative moment within the sector before GAI implementation continues further. GAI tools are developing rapidly, and keeping up with the multitude of uses and repercussions of them is a challenge.
Assessment – an unavoidable issue
Unsurprisingly, student assessments are a key point of discussion, and assessment practices in the face of GAI usage are highlighted as problematic. Staff are concerned that GAI is often misused by students and report having to look out for AI use in assessment work.
There is a common notion, repeatedly voiced, that student assessments need to be re-evaluated, and that GAI is a driver for change. There is a call for new assessment styles that are not only less vulnerable to GAI and reduce the opportunity for academic dishonesty, but more authentically reflect the abilities of students.
Despite the expressed need for vigilance around assessments and student use of GAI, there are many negative opinions (from learning technologists in particular) about detection tools. Some participants believe them to be ‘unreliable’ and ‘lagging far behind’, unable to keep up with GAI usage – this set of views further supports the need for assessment adaptation.
There is one aspect of assessment where GAI is discussed positively. It is proving useful for designing and developing assessments, as well as other learning & teaching resources. Participants report using it to generate quizzes, multiple choice question sets, refining questions and in some cases, marking and feedback.
The key roles GAI is being used for
The survey responses unpacked the role GAI is playing in the working life of the survey participants. In addition to assessment design, the survey highlighted some common use cases for GAI and key benefits of using it.
Some responders use GAI tools to feel more efficient, reduce workload and save time and effort, particularly on routine tasks. In some cases, these savings allow them to put more energy into other areas of their role.
One of the most tangible ways GAI is seen to improve efficiency, is in the role of starting point. Frequent references are made to GAI being used to remove blockers at the beginning of tasks (often involving blank page syndrome when writing). Many participants (learning technologists in particular) speak of GAI helping them make a start or to beat procrastination. GAI suggestions provide the leap required to clear hurdles.
Similarly, there is also evident regular use of GAI for generating and developing ideas and for tasks around planning and creating outlines, for example, to create a structure for a workshops or training session.
Simplification of information is another common application of GAI tools. Some participants have used it to summarize straightforward data sources, such as journal articles, or notes from meetings and discussions. Others utilize these tools for more intricate tasks related to their research or to help them swiftly grasp complex issues.
Some participants are turning to GAI to create and edit images, from quick samples and accessible images generated in line with guidance, to photo-realistic pictures without the need of extensive graphic design expertise. Image creation tools are aiding the development of learning resources, YouTube thumbnails, presentations and blog posts.
Writing and editing are frequently cited as a key reason to use GAI tools. Many survey participants indicated they used GAI tools for drafting email and work-related communications. They describe doing this to enhance their professionalism of the text and notably, to minimize the expression of a strong human emotion within it. This is described as happening in situations where the user is experiencing heightened emotions and trying to phrase their message in a more courteous manner to reduce the potential for conflict.
Several respondents explicitly indicated that their interactions with GAI tools resembled collaboration with a colleague, such as the exchange of ideas and the posing of non-critical inquiries. A sense of being in human conversation is described by some. Others noted using GAI for feedback on their concepts, and for guidance and consultation regarding their projects.
A small number of responses mentioned using GAI to improve accessibility. This also included discussion around frustrations around economic factors blocking the use of GAI tools as assistive technology – noting costs as a ‘disability tax’.
Building an understanding, encouraging an embrace
As mentioned earlier in the blog, there is some disparity around attitudes towards GAI from individual responders, and this is also true when it comes to uptake of the tools. While some respondents have found no need or reason to use the tools, and others feel that the tools are not good enough to use, many positively reported GAI being part of their daily activities. A significant number are going a step further – not just using the tools for ‘everything’ but promoting and encouraging a wider ‘embrace’ of them by colleagues.
Many survey participants feel that understanding the tools is a substantial task but one worth spending time on. A desire to grasp how the tools work, what they are capable of, and what their limitations are is a common driver to use them. Significant time is being dedicated to comparing the strengths and weaknesses of different tools and using them in an experimental way to evaluate how they can enhance tasks. Some are doing this for their own benefit, but many wish to cascade their knowledge to others, and use this experimentation to create training content and develop guidance.
A common experience described is that the greater the usage of the tools, the higher the confidence in using them; the greater the frequency of use; and the more use cases became apparent – simply, usage encourages more usage. Some of the ‘super users’ are encouraging others to make a start with the tools in the belief they will ‘embrace’ them and their use of them will grow too.
One motivator for this is the belief that it’s imperative to harness the use of GAI in HE to remain relevant. There is also a strong sense that it is crucial for education institutions to prepare students for a workplace that will be heavily shaped by these technologies.
Concerns and blockers remain
Of course, not all views expressed are wholly positive and not all responders are calling for an embrace of GAI technologies. A range of concerns about their use are raised.
It is commonly, and loudly, put forward that the quality and reliability of generative AI capabilities and its outputs are not to be trusted, with one responder going as far as referring to GAI outputs as ‘nonsense’ and ‘vacuous hogwash’. There is consensus that outputs need to be critiqued, and that staff need a robust understanding of what quality to expect from AI-produced content. Some respondents are worried that their colleagues fail to thoroughly examine the outputs.
There are also worries around intellectual property and a need to be ‘vigilant of the copyright issues’ – in some instances these objections are significant enough that GAI use is limited. There is an appetite for approved integrated tools that tie in with policy to be put in place, to alleviate these concerns, as well as others around data privacy and security.
The survey also identified strong concerns about the high costs associated with generative AI usage and the potential for impact on the digital divide. There is a fear that because some institutions may not have access to certain tools due to funding, this inequality may impact their ability to support students as well as other institutions – as earlier noted, it is felt this may be a particular issue for students for whom GAI tools play a role as assistive technology.
Another belief zealously voiced is that GAI use reduces creativity and stifles individuality, with one responder poignantly lamenting the loss of ‘beautiful, weird and unique voices’.
The survey revealed the impact GAI is having on staff skills. Some participants feel they need to enhance their current skillsets to be able to effectively use GAI, for example learning to write adequate prompts. Others believe using GAI will allow them to develop skills such as coding. In contrast, some responses describe fears that using GAI could contribute to their own ‘deskilling’. Others go as far to say that GAI may erase the need for their role in the future.
While there are concerns discussed, overall, the number of negative opinions raised is limited – but of those that do exist, some are powerful in their objection. One respondent highlighted discrimination by GAI stating ”It’s racist. It’s ableist. It supports genocide”.
As well as survey participants raising their own concerns, fears around generative AI are reported on behalf of others – responses shared that there is currently panic within the institutions about GAI and describe ‘intimidated’ colleagues that ‘lack confidence’ to engage.
Thoughts about the future
It is in thoughts around what is likely to happen in the future where widespread agreement from the survey participants is most apparent. There are three sentiments repeated frequently.
One – a belief that GAI tools are going to become more pervasive in the future.
Two – a belief that GAI tools are going to become more capable in the future.
Three – a belief that GAI tools will become trusted to be accurate in the future, and then use of them will grow.
Do you agree?