Closing the gap: addressing socioeconomic disparities in AI education access

A digital divide in students’ use of generative AI tools may be emerging, according to new research by HEPI and UCAS.

In this blog we delve into the survey findings, highlighting the importance of equitable access to AI in education and suggesting practical approaches to bridge the gap.

Students’ use of AI

The use of generative AI tools by higher education students is now widespread, following the public release of ChatGPT on 30 November 2022.

In the first UK-wide survey since its launch, HEPI and UCAS polled 1,250 undergraduate students to discover their attitudes towards generative AI tools.*

The research found that more than half of students (53%) have used generative AI to help them with assessments. The most common use is as an ‘AI private tutor’ (36%), helping to explain concepts.

AI digital divide

However, the survey findings also indicate an emerging digital divide, with some students benefitting more from generative AI tools than others.

The research uncovered that 58% of students from the most privileged backgrounds use generative AI for assessments, compared with just 51% from the least privileged backgrounds.

Further, students from the most privileged backgrounds were much less confident of their institution’s ability to spot AI use in an assessed piece of work than other students.

This suggests that students from the most privileged backgrounds may have greater access to generative AI tools and more confidence in using them without detection.

Equity of access

If generative AI tools have the potential to enhance education when used appropriately, then it’s crucial that all students can access these resources to ensure that none are unfairly disadvantaged.

In addition, access to AI tools will be important beyond higher education. Nearly three-quarters of students (73%) expect to use AI after university, according to the research, with common uses anticipated to be translating, enhancing and summarising text.

Again, the survey findings reveal a disparity: the most privileged students were much more likely to say they expect to use generative AI tools in the future than other students.

As before, this implies a greater familiarity with and confidence in using AI tools, which may result in the most privileged students being better prepared for employment in an AI-driven world.

Barriers to access

So, what are the barriers limiting the use of generative AI tools by some students?

The cost of subscription is a significant factor: currently access to ChatGPT-3.5 is free, but access to the more advanced ChatGPT-4 costs $20 (around £15) a month. It doesn’t sound like much in isolation, but these charges must be set in the context of the cost-of-living crisis.

The Student Money Survey 2023 found that the average student’s monthly costs have increased by 17% since 2022. Most students (82%) said they worry about making ends meet, while more than half (54%) had thought about dropping out because of financial pressures.

Digital and data poverty

Yet, unequal access to AI tools is not causing a digital divide – it may be widening an existing divide and exacerbating factors that already hamper inclusive learning experiences.

Some generative AI tools are free, but that’s not much help if you don’t have a laptop, access to stable broadband, or a suitable place to study, as we’ve explored previously.

Jisc’s Student Digital Experience Insights Survey 2022/23 found that over a quarter of higher education students (27%) had no suitable device. More than half (54%) struggled with Wi-Fi connection, while about a third (34%) worried about mobile data costs. A similar percentage of students (36%) had no private area to work, and around a fifth (19%) had no safe area to work.

As digital transformation in universities gathers pace, and more learning moves online, there’s a real risk of students from the least privileged backgrounds falling behind. So, what can we do?

Institutional provision

Three-in-ten students (30%) think institutions should provide AI tools, according to the HEPI/UCAS research, yet only one-in-11 students (9%) say they currently do so.

The report recommends that ‘institutions should provide AI tools for those who cannot afford them when they have been identified as benefitting learning’ to prevent the digital divide from growing.

Delivering on this need, we have developed a suite of cutting-edge, AI-powered study tools enabling students to summarise, create study notes, generate Q&A and translate text into over 100 languages, all applied to trusted, institution-approved content.

Bridging the digital divide

Yet, as we have seen, providing paid-for access to AI tools alone will not be enough to bridge the digital divide for some students.

Fortunately, there are many examples of good practice where universities are actively addressing digital poverty.

Most university libraries operate a short-term laptop loan service, while numerous institutions offer hardship funds to enable students to access a device for the duration of their studies.

The University of Manchester has developed a flexible learning strategy, encompassing equitable access to hardware and software, a seamless digital infrastructure and improved study spaces.

Several universities, together with the Digital Poverty Alliance and British Computer Society, have created a Digital Inclusion Manifesto and Toolkit to address digital poverty across the sector.

Enhancing AI literacy

We discussed students’ digital literacy in a previous blog. It remains an ongoing issue and, since the advent of generative AI tools, has been compounded by the need for AI literacy.

This skills element is also impacted by the digital divide. As we have seen, the most privileged students not only use generative AI tools more, but also express more confidence in using them.

The report recommends that institutions should teach students how to use generative AI ‘appropriately and effectively’, where it has been identified having benefits for learning, aligning with the Russell Group’s principles on the use of AI in education.

Again, there are plenty of examples of good practice by institutions in promoting AI literacy, such as guides or kits for students, and online sessions for educators. As AI tools become more embedded in higher education, the body of support materials will also grow.

The way forward

The emergence of a digital divide in the use of generative AI tools poses significant challenges for fair access at a time when entry rates of the most disadvantaged students are increasing.

Addressing this issue will require a multifaceted approach, and finding the means to implement strategies may not be easy when many universities are already facing financial challenges.

Nevertheless, institutions must take proactive measures now to ensure all students have equitable access to resources to equip every learner for success in an AI-driven future.

To discover how Kortext can support learning at your institution, tap here.

*Kortext provided some financial support for this research. However, full editorial control was retained by HEPI.