The Moment the Sector Has Been Avoiding
Article eight in The Proctoring Question series. Over seven articles, we have looked at how to make proctoring fairer, more secure and more defensible. This one asks whether we have been asking proctoring to carry too much.
The Debate Needs to Change
For years, the proctoring debate has been framed around a simple choice:
Remote or centre-based. Proctored or unproctored. Digital or paper.
But that debate is quickly becoming obsolete. Generative AI, second devices, smart watches, AI-enabled glasses and increasingly sophisticated identity spoofing have fundamentally changed the assessment environment.
Over the last year, I have witnessed a growing clamour in some quarters to return to the days of pen-and-paper assessments. Some organisations have grown tired of surveillance-heavy models, while others are struggling to keep up with evolving technology.
But the exam hall was never a fortress. Anyone who has been around for long enough will remember the sleepless nights, the exam day panics and the occasional disasters that we had to contend with. Going backwards doesn’t remove risk. It just takes us back to risks we learned to tolerate.
We are faced with a choice. We can respond by building ever-higher walls around the traditional exam. More monitoring. More lockdown. More detection. More surveillance.
Or we can ask whether we are really trying to protect the integrity of our assessments, or simply trying to protect assessment models that technology has begun to expose.
The Arms Race We Cannot Win
The proctoring model that we introduced during the pandemic served its purpose. It provided a solution that allowed us all to keep the lights on, and our assessments running. Without it, many of our organisations might not have survived.
But the world has changed. The reality is that surveillance-led proctoring eventually reaches a point of diminishing returns.
Every new threat creates pressure for us to add another layer of control. Second devices lead to room scans. AI tools lead to lockdown browsers. Smart watches lead to additional monitoring. AI-enabled glasses will inevitably produce another technological response.
But where does it end?
If every advance in technology requires another layer of surveillance, we have not created a sustainable integrity model. We have created an arms race. And it is an arms race that we simply cannot win.
The alternative is not to abandon security. It is to stop asking security technology to compensate for weaknesses in our assessment design.
Better Design, Better Integrity
Where stakes are high, there will always be pressure to cheat. But, as a sector, we have made a fundamental error in assuming that more surveillance equals more integrity.
For too long, we have treated assessment security as something that happens around the assessment. Lock the browser. Verify the candidate. Monitor the room. Record the session. Flag suspicious behaviour.
All of those controls can have value. But perhaps the most powerful security control sits inside the assessment itself.
If an assessment can be passed simply by retrieving information that an AI system can produce in seconds, the problem is probably not that our proctoring is insufficient. The problem may be that the assessment is asking for the wrong evidence.
Good assessment design can be a much more sustainable way to strengthen integrity. We cannot eliminate cheating, but we can make it less likely to succeed. Authentic, contextualised, process-oriented assessments make cheating harder to carry out and less likely to result in a candidate passing an exam without displaying genuine knowledge or competency.
Better design will mean different things in different contexts. It might mean allowing candidates to use information but requiring them to apply it. It might mean asking candidates to explain decisions rather than simply provide answers. It could mean gathering evidence across several activities rather than relying entirely on one high-stakes event.
There is no single model that will work for everyone. Nor does there need to be. But the principle stays the same: if an answer is easy to fake, ask for evidence that isn’t.
This is not an argument for replacing proctoring with assessment design. It is an argument for stopping proctoring from carrying a burden that better assessment design should be helping to carry.
Resilience Means Retaining Choices
One thing that I rarely hear in proctoring conversations is discussion around how so many of us ended up using it in the first place. COVID-19 exposed how vulnerable assessment models built around a single location and a single mode of delivery could be. We adapted because we had no choice.
The next major disruption may not be a pandemic. It could be geopolitical conflict, energy disruption, infrastructure failure, cyberattack or something we have not yet imagined. We cannot know what the next major disruption will be, or when it will happen. But assuming there will never be another one would be a dangerous basis for assessment strategy.
And if we want to remain resilient, we need to retain all of our options. An assessment organisation capable of delivering securely across centres, remote environments and hybrid models has options when circumstances change. One that has deliberately retreated to a single delivery model does not.
We all endured a lot of pain during the emergency adaptations of the pandemic. We need to learn from that experience, not dismantle what we built because that particular emergency has passed.
What About Proctoring?
Proctoring still has an integral role to play in the future of assessment delivery. But we must stop trying to measure its success by how much of the assessment environment we can force it to monitor.
Its future may be narrower, more targeted and less visible. The answer is not necessarily more aggressive monitoring, but strong controls around the moments that genuinely require them, and considerably less intervention where they do not.
AI can help identify anomalies and prioritise review. Humans can interpret context, exercise judgement and make consequential decisions.
We need to remember that the objective should not be maximum surveillance. It should be sufficient assurance.
A Final Thought
The future of proctoring will not be determined by whether we pick remote assessment over the exam hall, or whether we pick digital over paper. Those are yesterday’s arguments.
The real challenge is whether we continue adding technology around assessments in an attempt to protect them from a changing world, or whether we become brave enough to change the assessments themselves.
Proctoring will remain part of that future. Sometimes we will need strong controls. Sometimes human observation. Sometimes automated detection. Sometimes very little monitoring at all.
The question is not “How much proctoring can we apply?” but “How much assurance does this assessment actually need?”
It is time for us to seriously address the question that we have been avoiding for so long: is it only our proctoring model that needs to change, or is it our assessment model too?
We can continue to fuel the arms race between cheating technology and surveillance technology. Or we can design assessments that provide better evidence in the first place.
The full series
The Proctoring Question began with Designing Trust: The Next Chapter of Remote Proctoring, which set out eight themes that decide whether remote proctoring earns and keeps trust.
- Fairness across delivery channels
- Reviewer workload and flag fatigue
- How Secure Is Secure Enough?
- Choosing the right proctoring model
- The Candidate Experience Is Part of Assessment Quality
- Access and Equity Are Not the Same
- Who Decides What a Flag Means?
- The Future of Proctoring (this article)
Coming next week: [downloadable title], [one line on what it gives the reader].
More surveillance is not the only answer. Speak to one of our experts about how much assurance your assessments really need, and how much of it better design can provide.
