Reporting of regulatory fines linked to errors in digital assessment has prompted questions about the role of technology in high-stakes assessment. For some, such incidents are presented as evidence that the sector should slow down, or even reconsider, the move towards digital assessment altogether.
That is the wrong conclusion.
When assessment errors affect learners, the consequences can be significant. Assessment organisations have a responsibility to ensure robust quality assurance processes are in place, and regulators are right to intervene when standards fall short. But it is important to recognise that assessment errors are not unique to digital delivery, automated marking or AI-assisted systems. They have existed throughout the history of assessment, including in entirely paper-based environments.
Assessment errors are not new
Every year, thousands of exam results are reviewed following requests from schools and candidates who believe a mark or grade may be incorrect.
Ofqual's latest statistics show that, in summer 2025, 4.6% of all GCSE, AS and A level grades awarded in England were challenged through a review process. Of those challenged grades, around a quarter were amended, resulting in 72,880 grade changes. Overall, 1.1% of all grades awarded were subsequently changed following review. The figures are remarkably consistent year after year and reflect a reality that assessment professionals have long understood: human judgement is not perfectly consistent. Different trained markers can legitimately interpret responses differently, particularly in subjects involving extended writing and professional judgement.
The existence of reviews, remarks and appeals is therefore not evidence of failure. It is evidence that assessment systems recognise the possibility of error and include mechanisms to identify and correct it.
The real question is how errors are managed
The recent focus on digital assessment risks creating a false distinction between "technology errors" and "human errors".
Assessment systems can fail for many reasons. Question papers can contain mistakes. Mark schemes can be interpreted differently. Administrative processes can go wrong. Data can be entered incorrectly. Technology can fail. Human beings can make mistakes.
The important question is not whether errors are possible. It is how quickly they are identified, how effectively they are corrected and what controls exist to prevent them occurring in the first place.
By that measure, digital assessment often provides advantages rather than disadvantages.
Digital assessment creates new opportunities for quality assurance
One of the less discussed benefits of digital assessment is the visibility it provides.
Paper-based marking processes have historically relied on sampling, moderation and post-results review processes to identify inconsistencies. Digital systems can go much further. Online marking platforms can monitor marker performance in real time, identify unusual patterns, introduce quality-control responses, track decision making and provide detailed audit trails. Issues can often be identified before results are released rather than afterwards.
Increasingly, assessment organisations are also exploring how AI can be used not as a replacement for human judgement but as an additional layer of quality assurance.
At the eAA’s recent International e-Assessment Conference "The Trust Imperative" was the central theme and many of the discussions focused on exactly this challenge: how technology can be used to strengthen trust, transparency and accountability in assessment. Across conference presentations, research papers and award finalist case studies, a consistent theme emerged. The most mature implementations are not seeking to remove human oversight. Instead, they are introducing multiple layers of verification.
Examples include:
This is a very different proposition from the simplistic narrative that "AI marks exams".
Building trust through evidence
Perhaps the most important lesson from recent events is that trust in assessment should be based on evidence, not assumptions. The assessment community has spent decades improving reliability through standardisation, moderation, quality assurance and continuous review. Digital technologies provide new tools to support those objectives, not replace them.
The conversations taking place across the assessment sector today are increasingly focused on how to combine the strengths of technology with professional expertise. The goal is not automation for its own sake. The goal is more consistent, more transparent and more resilient assessment.
Assessment organisations will continue to make mistakes. Paper-based systems make mistakes. Digital systems make mistakes. Human markers make mistakes. Technology can make mistakes. The challenge for the sector is not to eliminate all possibility of error. It is to build systems that detect problems quickly, learn from them and continuously improve.
That is ultimately where trust comes from: not from the absence of technology, but from the strength of the governance, controls and accountability that sit behind it.
About the e-Assessment Association
As the global community for digital assessment, the e-Assessment Association brings together exam boards, regulators, technology providers, educators and employers to define best practice, share evidence and build trust in digital assessment.
Through collaboration, evidence and thought leadership, the eAA is helping shape a future where assessment is more accurate, inclusive and trusted for learners, institutions and employers worldwide.
Individuals can join the eAA for free. Find out more here.
Join the global community advancing e-assessment through innovation, research, and collaboration.
Keep informed
This site uses cookies to monitor site performance and provide a mode responsive and personalised experience. You must agree to our use of certain cookies. For more information on how we use and manage cookies, please read our Privacy Policy.