We are at a critical juncture in higher education, where equity in AI holds the potential to broaden access to learning--or widen gaps.

Defining a path to equitable AI in higher education

AI could be a powerful tool for equity in learning--or could represent a technological advancement that further widens the digital literacy gap

Key points:

In today’s information-driven world, the ability to access, aggregate, analyze, synthesize, and otherwise use vast amounts of information and access knowledge on demand increasingly makes data literacy an absolute necessity. The potential of AI to dramatically transform access and attainment through personalized learning at scale, as well as in providing agency to every learner in following paths to better success, is extremely attractive.

However, the exacerbation of the digital divide (often described as the gap between those who have access to technology, the internet, and digital resources, as well as the skills to use these, and those who do not) with the potential addition of systemic bias and gaps in AI literacy are concerning.

While the topic has long been discussed in academic, legislative, and policy circles, the extent and impact of the gap were perhaps only brought to the forefront, and comprehensively understood, by the COVID pandemic and the consequent need to largely move education online.

Despite significant efforts over the past three years to bridge the gap and address the divide, recent reports suggest that as many as 23 percent of Americans do not have access to a broadband connection at home, and as many as 33 percent do not have speeds sufficient for efficient video conferencing, let alone online/digital learning with streaming and immersion. Forty percent of schools still lack broadband access and the median bandwidth across the nation is just 1.25 Mbps, significantly lower than the US FCC minimum threshold for broadband connections of 25 Mbps for download and 3 Mbps for upload.

The disparity in awareness, availability, affordability, adoption, and use of digital technologies/tools at institutions of higher education, and thus by learners, profoundly disadvantages some not only decreasing opportunities for learning but placing barriers in pathways to success and improved socioeconomic mobility. These disparities disproportionately affect the very segments of society who would benefit the most.

With the increasing use and adoption of AI in the workplace, it is imperative that institutions of higher education prepare students for success in this fast-changing world. In large part, this necessitates not only that the existing digital divide is not widened, but rather that steps be taken to close that gap while also attaining equity in AI. It is thus critical that the higher education sector, as a group, define equity as related to AI with a view of identifying critical facets that could contribute to furthering inequity and disparity, develop mechanisms and paths towards addressing them, and advocate for the resources and policies to make this possible.

It is perhaps tempting to consider that AI equity could be simply defined in the context of higher education as learners having equal access to technology, devices, the internet, and experts to assist in enabling their use. However, the reality is a bit more complex, because equal access by itself does not translate to equity of outcomes, which must be considered in the context of rapidly shifting workforce needs, which will necessitate AI and digital literacy as prerequisites for a growing number of paths to socio-economic mobility. In general, the AI divide, which must be overcome to ensure equity as related to AI in higher education, could be thought of in terms of four levels.

Level 1: Access to technology/tools:  This fundamental level builds on the precepts of availability, affordability, and adoption of AI in the context of learning and preparation for success of students in the workforce. At a minimum, this encompasses students having affordable access to (a) reliable, high speed broadband connectivity, (b) devices on which AI programs and platforms are accessed and used, and (c) the basic AI tools used for learning and ensuring adequate preparation for the workforce and life in an information driven digital world. It must be emphasized that lack of these presents a fundamental barrier to attaining capacity for full participation in society and for ensuring socioeconomic mobility through higher education. While access is important, the affordability of the devices, programs and connectivity will need sustained funding and partnerships outside higher ed.

Level 2: Access to training and expertise: Just as with the digital divide, access at level 1 is meaningless without the ability to access and use expertise necessary to adequately train both faculty and staff as well as the ability of institutions to adapt tools and platforms to the specific contexts in which they will be used and for the diverse populations being served. Because a fundamental advantage of AI is the ability to specialize and personalize paths and interactions it is critical that differences in institution based on mission and segments of populations served are adequately represented in the design of AI tools including both data used for training of the AI tools and in the specifics of the support provided to the student. In similar fashion the appropriate adaptation of language and socio-cultural context is critical during design without which historically underserved/unserved populations would only be further disadvantaged. Ensuring that key constituencies have “a seat at the table” and that their voices are heard during design and validation of tools will be crucial to achieving this. It is also critical that there be greater, and closer, collaborations/partnerships between technology companies and institutions of higher education so that there is a better understanding and synthesis of needs and capabilities during design of tools with a goal of better serving learners. Often opportunities in ed tech are missed because of a lack of conversation about needs from academia and about what is possible from the technology sector.

Level 3: Access to resources to attain AI literacy: Attaining this level of literacy, and hence AI equity, will require having the ability to actively use AI technology to analyze, develop, and create rather than just passive incorporation of AI into aspects such as assignments, “black box” analytics, or as passive consumers of content provided by others which may not meet the needs of the specific institution/population segment. This requires knowledge about data sharing and governance and necessitates an elevated level of agency on the part of the specific institution (and hence for the faculty and students) rather than just at the vendor level. Effectively then, this level of equity requires the formation of unique partnerships focused on outcomes and includes aspects such as compliance and ethics of representation and use of data/information. A number of these topics have nuances and significance because of vendor supplied platforms using AI without the institution being aware of it, and the implications thereof, as well as due to the intricacies and complexities of integrating a range of platforms together while ensuring integrity of data and the ethics of its use.

Level 4: The ability to apply these to gain equity in socioeconomic mobility: The highest level of equity is based on ensuring equal opportunity from use of the tools and must necessarily address critical aspects such as input bias (implicit, historical/intrinsic, unintended, or due to lack of inclusion of appropriately representative data sets) and that of output, as well as of emerging challenges such as copyright and IP ownership. This impacts, at a profound level, the ability of institutions to truly meet the needs of learners they serve in ensuring that talent meets opportunity for socioeconomic mobility. Thus, achieving transparency and accountability are essential to overcome this level and will require not just collaborations between vendors and institutions of higher education, but thoughtful consideration of key aspects in policy at the federal, state, and accreditation levels, requiring the convening of focused councils centered around ensuring student success. In many cases policy, by itself, will not suffice and there need to be thoughtful, and open, discussions about the implications of AI and its use as an integral part of learning and the operation of institutions of higher education. While the challenges may never before have been as hard, complex, and intertwined, the benefits that could accrue to our students and to the addressing of complex issues here to fore considered too difficult to resolve at scale could be transformational.

The four levels could be generically considered as overlapping with the three basic characteristics of digital access, digital design, and digital use mentioned in the 2024 National Educational Technology Plan. The overlap in characteristics of digital equity and those of AI equity demonstrates the interconnected nature and further emphasizes the additional impact of AI as related to inequity if not approached from an appropriate basis. It is perhaps, as important, to consider these levels in terms of the desired output with levels 1 and 2 contributing to the further democratization of knowledge and learning, levels 2 and 3 focusing on operationalizing the critical facets of this, and level 4 assuring that the focus is on moving away from the outmoded “one size fits all” assembly line modality to one that enables personalization as a means of further access and attainment in higher education.

In this context, as related to higher ed, AI and digital technologies cannot be considered in isolation but must be thought of as interdependent, with the sum being greater than their parts, especially as related to the negative impact potential.

While there are significant other concerns associated with the use of AI tools for learning, especially as related to data privacy, cybersecurity, intellectual property and copyright, and even depersonalization of learning if automation is taken to an extreme, at a minimum the above referenced levels need to be foremost in any discussion related to implementation for an institution or higher ed sector, highlighting existing barriers to achieve digital equity, providing promising strategies to overcome these barriers, and identifying key action steps that can be taken by leadership.

Key among these will include the following:

  1. Development of partnerships between academia and vendors that provide appropriate focus to learning/tutoring/support tools being developed such that they meet existing, and future needs, enhancing pedagogy, enabling better and more nuanced learning through “discovery” and “enquiry” while also ensuring increased critical thinking and analysis, making real the effect of Bloom’s 2-sigma problem in enabling every learner to have personalized assistance that builds from strengths rather than deficits.
  2. Development of specific funding mechanisms through a combination of federal, state, and foundation-based support that not only decreases the digital divide but also enhances resources for institutions, individually and/or as groups, to have access to both the appropriate AI tools and the expertise to be able to actively use these. This will, in all probability, require both new funding streams and re-envisioning sources and formulae to ensure equity across the board, at least at the basic level (Levels 1-3 as described previously).
  3. Leadership by key national organizations which can serve as unbiased convenors in the creation of mechanisms to bring interested parties together and to ensure continued thoughtful discussion on these topics always placing the learner/student as the focus. Given the disparate systems, accreditation agencies, and other regulatory bodies such mechanisms will be critical to ensure that even if equity is achieved at one instant in time it does not get dissipated with the next set of technological advances or funding policies.
  4. Provision of mechanisms by which institutional and system leadership, including boards, are continuously updated on the potential and ramifications of AI and digital technologies ensuring that they are knowledgeable enough to provide wisdom through appropriate leadership and governance. The same must be assured with state legislatures and higher education boards, both of which could either close, or widen, the digital and AI divide, with profound, and long-lasting, implications.
  5. Mechanisms by which success stories (including methods/approaches followed and data) can be shared with full detail, as well as failures, again keeping the best interests of learners in mind. This will probably need the development of key Centers, at leading institutions of Higher Education with a mission of access rather than exclusion, funded through the Department of Education, the National Science Foundation, the National Institutes of Health, foundations committed to learning, and consortia of ed tech companies.

We are at a critical juncture, one where AI could be a powerful tool for equity in learning or could represent a technological advancement that further widens the digital literacy gap, and hence the disparity in reaching socioeconomic prosperity, in the 21st century. AI is a complex issue and the importance of equity in AI as related to learning and preparation of students for the future cannot be overstated. Institutions of higher education, and bodies committed to ensuring equity in learning as a means of talent meeting opportunity and enabling greater socioeconomic mobility do not have the luxury of sitting on the sidelines to see how the evolution of AI transforms society. The future demands the rapid evolution, and implementation, of thoughtful strategies and partnerships to ensure equity in learning access and attainment. Our students deserve nothing less than our full engagement and commitment at this critical juncture.

Sign up for our newsletter

Newsletter: Innovations in K12 Education
By submitting your information, you agree to our Terms & Conditions and Privacy Policy.

eSchool Media Contributors