As academic misconduct rises, employers face a hard question: can qualifications still guarantee capability in the workplace?

The gold standard of student assessment has long been the oral viva.

Ask a student to explain, justify and defend their work through conversation, and it quickly becomes clear whether they genuinely understand it. If every assessment could be accompanied by a viva, concerns about ChatGPT and AI-generated coursework would largely disappear.

The problem is scale.

Delivering oral assessment across every module, every assignment and every student is simply not practical for most universities.

That challenge is becoming increasingly urgent as artificial intelligence reshapes higher education.

Artificial intelligence is forcing universities, employers and policymakers to confront an uncomfortable question: how do we know that what a student submits reflects what they actually understand?

A Guardian investigation found nearly 7,000 confirmed cases of AI-related academic misconduct in UK universities during 2023–24, while separate research suggests 92% of students are now using generative AI tools in some form. (HEPI and Kortext Student Generative AI Survey 2025)

But the issue is bigger than misconduct.

It is about whether degrees can still be trusted as signals of capability. Universities are responsible for developing the talent pipeline that underpins the UK economy. If graduates leave with credentials but without the skills those credentials are supposed to represent, businesses face growing uncertainty about workforce readiness.

This is not simply an academic concern. Britain’s productivity, innovation capacity and global competitiveness depend on the credibility of its higher education system. Degrees are trusted because they are assumed to represent capability. If that trust weakens, the impact will extend far beyond the classroom.

Universities are already responding. Oxford has updated its policies around AI in assessment, Sussex is exploring “authentic assessment” approaches designed to reduce AI dependency, and Portsmouth has introduced guidance to help redesign coursework and examinations for the AI era. Some institutions are returning to in-person handwritten exams, while others are restricting AI use altogether.

Yet these responses only scratch the surface of a deeper challenge: how do we verify genuine understanding in a world where AI can generate polished answers in seconds?

Universities already know what good assessment looks like. The challenge is making it scalable.

That is where new approaches are emerging.

One of them is Integrevise, a platform designed to help universities verify genuine understanding at scale through AI-enabled oral assessment and learning intelligence tools.

The challenge dominated conversations at this year’s Times Higher Education Digital Universities Conference, where Integrevise was named Startup of the Year.

Discussions with Pro Vice-Chancellors, Heads of Digital Learning and academic leaders revealed a consistent concern: universities do not currently have a scalable way to verify genuine understanding in an AI-enabled world.

“Winning Startup of the Year was hugely encouraging, but the biggest takeaway from the conference was how consistent the conversations were. Universities know the assessment model needs to evolve. The question now is how to do that in a way that is scalable, credible and practical for staff and students alike.” – Kyle Poyer, Integrevise

The reality is that traditional written assessment models were not designed for the AI era. Universities now need approaches that test not just output, but comprehension, reasoning and the ability to apply knowledge under pressure.

Universities have always understood the value of oral assessment. The challenge has been making it practical at scale.

Integrevise addresses that challenge by automating oral assessment while also giving educators real-time insight into learning gaps, comprehension levels and student progress. By turning assessment into a source of actionable learning data, the platform helps universities strengthen confidence in qualifications while giving employers greater confidence in graduate capability.

Integrevise was created through the Cyber Innovation Hub Venture Building Programme, delivered in partnership with Cardiff University and the Alacrity Foundation. Unlike many startups that begin with a technology and search for a use case, Integrevise started with a clearly defined challenge: how can universities maintain trust in assessment and qualifications in the age of AI?

Through the programme, the team developed and validated an approach that combines scalable oral assessment with secure data handling and learning intelligence.

Early pilots with UK institutions suggest the approach can reduce staff workload, improve confidence in assessment outcomes and help institutions adapt more effectively to the realities of AI-enabled learning.

The AI era is already reshaping education. Universities cannot preserve the value of degrees by relying on assessment models built for a pre-AI world. They must treat AI not simply as a threat to control, but as a catalyst to rethink how learning is tested, verified and trusted.

Recognition at the Times Higher Education Digital Universities Conference reflects the growing urgency of the challenge Integrevise set out to solve and the need for new approaches to assessment in the AI era.

Stories like this show what can happen when cyber security research is connected to real-world challenges, commercial insight and the right support.

CyberInvent gives future researchers the opportunity to follow that path through a fully funded PhD at Cardiff University, industry-led cyber challenges and Cyber Innovation Hub’s Venture Building Programme, where ideas can be tested, validated and developed into real-world ventures.

If you have the curiosity, technical potential and ambition to solve a cyber security problem that matters, CyberInvent could be your next step.