Across UK education, disabled students and staff are facing greater barriers at precisely the point where the support structures meant to remove them are being cut back.
Reductions to Disabled Students’ Allowance (DSA), delays to Access to Work (ATW), and rising expectations from regulators have left universities expected to do more with fewer tools, shorter timelines, and significantly less certainty than their school and FE counterparts.
At the heart of this is a simple but urgent problem – equality impact assessments (EIAs) are being used as records of compliance rather than as tools for co-production and anticipatory design.
When EIAs are completed after decisions have already been made, without meaningful participation from those most affected, they can’t fulfil their purpose as safeguards. Making EIAs genuinely participatory and forward-looking isn’t optional – it’s essential if institutions are going to prevent harm rather than merely document it.
Behind every policy shift is a person who must now work harder to access what others take for granted, and during 2025 we’ve seen just how fragile our systems can be at the exact points where people most need them to hold.
Decisions made far from classrooms, clinical placements, and apprenticeship settings ripple into daily realities in ways that are rarely captured in official paperwork – least of all in EIAs, which are too often completed late, procedurally, and without meaningful participation before decisions are taken.
If the sector is serious about justice, belonging, and safety, EIAs can’t continue as rituals of minimal compliance. As critiques of recent departmental EIAs have shown, proceduralism alone doesn’t prevent inequity – it can entrench it.
A more pressing question follows – what might higher education look like if EIAs were treated not as forms to complete, but as tools to think with? As ways of building foresight, collaboration, and harm prevention into decisions, designed with the people they most affect.
What needs to change
EIAs retain enormous potential, but only if the sector is willing to rethink their purpose and practice. Turning them from administrative processes into something more useful begins with four shifts.
First, they must be co-produced – developed with the people they affect rather than applied to them, including disabled learners, disabled staff, practitioners, accessibility specialists, and assistive technology experts, whose participation ensures that assessments reflect lived experience rather than abstract assumptions.
Second, the process itself needs to become more human and participatory, asking questions that go beyond compliance – what does this decision feel like in practice, who carries the greatest burden, who benefits and who is excluded, and what forms of risk emerge and for whom. Lived experience isn’t anecdotal – it’s vital evidence about how systems actually function.
Third, EIAs need intersectional and power-aware analysis that recognises disabled communities aren’t homogenous. Experiences differ by gender, race, socio-economic background, language, disability type, and more, and without attention to these dynamics, EIAs risk hiding harm rather than identifying it.
Fourth, they should be anchored in universal design and anticipatory duty. Accessibility should be treated as foundational infrastructure, not something bolted on late in the process – EIAs belong at the beginning of conversations, in design briefs, early governance papers, and programme planning, and should be regularly revisited as projects evolve. This aligns with wider sector expectations around anticipatory duty, digital accessibility, and inclusive assessment.
Beyond compliance
In many institutions, EIAs have drifted from anticipatory safeguarding to administrative box-ticking. They’re often initiated once decisions have already taken shape, circulated for signature, and archived as evidence of due process – and in this model, disabled people become abstract categories rather than contributors whose insights could shed light on both risk and opportunity.
The 2025 Department for Education (DfE) decision to withdraw DSA-funded assistive software offers a clear example. On paper, the EIA followed the Public Sector Equality Duty (PSED) process, but the assessment relied on narrow measurement tables, ambiguous definitions of “free” provision, and assumptions about the equivalence of mainstream tools – and it didn’t engage disabled learners, practitioners, or assistive technology specialists. The result was a procedurally correct document that failed to anticipate disproportionate harm.
This isn’t a one-off – it reflects a broader pattern in which EIAs that prioritise procedural completion over substantive engagement can’t meet the anticipatory duty established in law, and risk recording inequality rather than preventing it.
Ripple effects and digital gaps
Higher education is now experiencing the consequences of EIAs that meet procedural requirements while missing human impact. The effects of DSA reforms, ATW delays, and increasingly complex benefits systems fall hardest on the very people who rely on support to participate equitably – for some, these delays and gaps translate into increased cognitive load, disrupted learning, and avoidable health impacts, while for institutions they contribute to higher risk, increased casework, and diminished confidence in systems intended to protect learners.
Within universities, similar patterns play out in assessment practices that remain inaccessible by design, digital environments that don’t consistently meet WCAG 2.1 AA standards, and policies that rely on “exceptions” rather than anticipatory infrastructure.
These issues disproportionately affect neurodivergent learners, disabled staff, apprentices who depend on workplace accessibility, and students for whom English is an additional language.
People, not paperwork
The 2024 High Court ruling in University of Bristol v Abrahart made the anticipatory duty explicit – institutions must identify and mitigate foreseeable risks before harm occurs. Current EIA practice, with its inconsistent timing, variable quality, and limited participation, doesn’t always meet this expectation.
When EIAs fail to account for within-group differences, socio-economic disparity, linguistic diversity, neurodivergence, or long-term health conditions, they can’t reliably protect the people they’re designed to safeguard. Procedural compliance alone is simply too blunt an instrument for the complexity involved.
A participatory, anticipatory EIA process would let institutions identify risks early and adjust direction before harm occurs, test proposals through diverse user scenarios, understand long-term implications rather than just short-term mitigations, build accessibility and belonging into organisational design, reduce grievances and reactive casework, and strengthen trust among students, staff, and partners.
This isn’t an argument for more bureaucracy – it’s a redistribution of effort. When accessibility is embedded from the outset, organisations spend less time responding to crises and more time building sustainable equity.
The shift also aligns with broader system changes. As responsibilities move between government, regulators, and institutions, universities can’t rely solely on external support. A stronger internal approach to accessibility and equity isn’t only morally and legally necessary – it’s operationally pragmatic.
Partnership and AI
A modern EIA approach can’t be developed by the higher education sector in isolation, because the challenges faced by disabled learners and staff sit at the intersection of education, health, employment, and social care policy.
Organisations such as the National Association of Disability Practitioners (NADP), British Assistive Technology Association (BATA), SUs, and professional bodies all have expertise and lived insight to contribute. Regulators – including the Office for Students and Ofsted – and key bodies such as the DfE and the Department for Work and Pensions (DWP) have a role in setting expectations that recognise participation, rigour, and anticipatory reasoning as markers of quality.
At a time when emerging technologies offer new opportunities for access and participation, the sector has a chance to create a coordinated, future-oriented approach to equity, and participatory EIAs can act as a bridge between policy, practice, and lived experience – supporting decisions that aren’t only defensible but fair.
Used responsibly, artificial intelligence has the potential to enhance accessibility in ways that were unthinkable even a few years ago. It can’t replace lived experience, specialist expertise, or co-produced design, but it can help create more inclusive environments – particularly for learners who don’t meet funding thresholds or whose support is disrupted by systemic delays.
A future in which no student is denied opportunity because of the limits of their support system is possible, but only if we take the necessary time and care to build purposeful AI use cases collaboratively, ethically, and with accessibility at their core.
What now
The sector can no longer afford to treat EIAs as administrative hurdles. The policy context has shifted, the risks have escalated, and the consequences of misjudged decisions have become visible.
Universities and partner organisations now have both an opportunity and a responsibility to rebuild EIAs as tools that genuinely protect people, guide ethical decision-making, and strengthen equity – and this requires cultural as well as procedural change, with a commitment to listening, to co-production, and to embedding accessibility as an organisational principle rather than an afterthought.
There’s also a hopeful dimension here. Used responsibly, AI can support accessibility in ways that complement specialist provision and lived expertise – and if universities, disability practitioners, disabled students, and policy leaders work together, we can shape AI not as a cost-saving measure but as a tool for dignity, independence, and equitable participation.
EIAs won’t solve every challenge, but if they’re reimagined as collaborative, human, and anticipatory, they can help prevent the harms that procedural compliance too often overlooks – and they can help institutions demonstrate not only that they followed a process, but that their processes are worthy of the people they serve.
