We talk to a lot of programs right now that are dealing with the same thing. AI tools have changed what cheating looks like, and most proctoring systems were built for a problem that’s no longer the only one on the table. The gap is showing up in appeals, in complaints, and in outcomes that are getting harder and harder to defend.

What’s actually happening out there

The classic image of cheating a second phone, a friend on the other side of the room, notes taped to a monitor is still real. But it’s not the conversation we’re having with most programs anymore. The harder stuff is subtler.

AI writing tools can produce a natural, well-reasoned answer in seconds. Paraphrasing tools can disguise lifted content well enough to pass similarity checks. And the behavioral signals that used to flag something suspicious eye movements, typing pace, browser switching don’t tell you much when the assistance is happening invisibly, in another tab or on another device entirely.

The problem isn’t that AI cheating is impossible to catch. It’s that catching it requires a level of context and judgment that an algorithm alone doesn’t have.

An algorithm can tell you something looked unusual. It can’t tell you what actually happened or whether it matters. That part still requires a person.

‍

Why a flag without a human isn't enough anymore

Automated proctoring was designed to catch visible, definable behaviors tab switching, phone use, an unauthorized person on screen. For those cases, it works. But AI-assisted cheating often leaves no visible trace at all.

So when an automated system flags something in this environment, what it's actually telling you is: something here didn't fit the expected pattern. That's a starting point, not a conclusion. The problem is when programs treat it like one.

A wrongful finding based on an automated flag isn't just uncomfortable it can seriously damage a student's record or a professional's career. And when that decision gets challenged, "the system flagged it" isn't a defensible answer. You need a record of what actually happened and a human judgment behind the decision.

{{post-stat-highlight}}

What it actually takes to be defensible

Defensibility isn't really about technology. It's about being able to look anyone in the eye a student, a candidate, a regulator, a board member and explain clearly why a decision was made and what it was based on.

In this environment, that takes three things:

• Human review on every flag, before any decision goes out. Not as an appeal process. Before the outcome is issued.
• Reviewers who are looking at AI-use patterns specifically, not just traditional integrity signals, and who understand the difference between something suspicious and something definitively wrong.
• A complete session record identity verification, monitoring log, what was flagged, what the reviewer saw, what was decided ready to export the moment someone asks.

What actually changes with human review

One question to sit with

If an outcome from one of your proctored exams was challenged today by a student, a candidate, an employer, an accreditor what could you actually put in front of them? An automated flag, or a complete human-reviewed record of what happened and why?

For programs where the result genuinely matters, that's not a rhetorical question. It's the one worth answering before you need to.

We built Integrity Advocate around exactly this human review on every flagged session, a complete audit trail for every outcome, and the kind of defensibility that holds up when someone actually pushes back.

{{post-cta}}

There is a difference between monitoring and surveillance. Most programs running online assessments haven’t thought carefully about which one they’re actually doing, and that gap is where trust starts to erode.

Monitoring confirms that the right person completed the assessment, stayed engaged, and followed the rules. Surveillance collects everything it can, flags anything it doesn’t recognize, and leaves the program to sort out what any of it means.

One produces a defensible result. The other produces liability.

If your proctoring tool accesses the file system of a personal device, stores recordings your team will never review, or generates hundreds of automated flags with no human judgment behind them, it’s not monitoring. It’s surveillance. And your learners know the difference.

What monitoring is actually supposed to do

Every program administering online assessments has the same core need: confirm identity, verify participation, and produce a result that holds up to scrutiny.

That’s it.

You’re not trying to catalog everything that happens in a candidate’s home. You’re not trying to build a behavioral profile. You’re trying to answer three questions:

1. Was this the right person?
2. Did they complete the assessment under appropriate conditions?
3. Can you stand behind the result if it’s challenged?

A well-designed proctoring approach answers all three. An over-engineered one tries to answer questions nobody asked, and in doing so, creates new problems. More data. More noise. More privacy exposure. And a test-taking experience that generates complaints before a single result is issued

Where surveillance starts

Privacy-first proctoring is not a softer version of security. It’s a more precise one.

It starts with a design principle: collect only what is necessary to confirm identity and verify engagement. Nothing beyond that scope. No recordings stored by default. No system-level device access. No behavioral data used outside the bounds of the specific assessment.

In practice, that looks like this:

No downloads or extensions. The assessment runs in a browser. No software installs, no hidden permissions, no access to personal files or background applications. A test taker can sit the exam on any device without IT involvement or configuration.

Proportionate data collection. Every feature and every data point collected should have a clear answer to the question: why does this make the result more defensible? If the answer is “it doesn’t,” it shouldn’t be collected. Surveillance produces volume. Monitoring produces clarity.

Data minimization by design. The only data collected is what’s needed to produce a defensible result: identity verification, participation confirmation, and a reviewable record of any flagged behavior. Sensitive data, including ID images, is deleted within 24 hours.

Compliance built in, not bolted on. GDPR, PIPEDA, FERPA, and SOC2 compliance aren’t checkbox items. They reflect a commitment to proportionate data collection, the kind of approach that holds up when regulators or institutional counsel start asking questions.

Human review is what makes the result defensible

Here’s what automated flags can’t tell you: whether a behavior was actually a problem.

An algorithm can detect that a second face appeared in the frame. It cannot tell you that the second face belongs to a child who walked in, not an accomplice. It can flag an unusual eye movement pattern. It cannot determine whether that pattern indicates academic dishonesty or a medical accommodation. It can detect a background sound. It cannot decide whether that sound matters.

Those determinations require a person.

At Integrity Advocate, every flagged session is reviewed by a trained human reviewer before any outcome is issued. Not as a premium tier. Not as an add-on. As the standard, at every price point.

That’s the difference between a defensible result and an automated flag you have to explain to a grievance committee.

Programs that rely on automated proctoring alone are producing outcomes they cannot fully stand behind. When a result is challenged, and eventually one will be, “the algorithm flagged it” is not a sufficient answer. A documented, human-reviewed decision is.

The question your program should be asking

Before your next contract renewal or platform evaluation, ask this: what does your current proctoring tool collect, and who reviews it before a result is issued?

If the answer is “a lot” and “no one,” that’s the surveillance model. It may feel thorough. It isn’t defensible.

Privacy-first online proctoring isn’t about doing less. It’s about doing exactly what’s necessary, with human judgment behind every decision that matters. That’s what produces results your program can stand behind. It’s what earns trust from your learners. And it’s what holds up when the stakes are highest.

Integrity doesn’t require surveillance. It requires precision.

{{post-cta}}

There is a meaningful difference between a proctoring system that catches violations after they happen and one designed so that violations are less likely to happen in the first place.

Most platforms focus on detection: flag the behavior, generate a report, let the program sort out what it means. The problem with that model is that it puts the burden on test takers, creates false positives, and produces results that are hard to stand behind when challenged.

Integrity Advocate takes a different approach, one grounded in a principle that applies equally well in physical security, public safety, and assessment integrity: the most effective systems deter problems before they occur, respond proportionally when they do, and produce evidence that holds up to scrutiny.

{{download-resource-cta}}

Deterrence First

The presence of proctoring changes behavior. Most test takers who know their session is being monitored by a real person, not just logged by an algorithm, approach the assessment the way it was intended. No intervention needed.

This is why the combination of visible monitoring and human review is more effective than automated detection alone. It is not just about catching the problem. It is about creating conditions where the problem is far less likely to arise.

Proportional Response

Not every behavioral signal warrants the same response. A proportional approach distinguishes between a minor anomaly and a genuine concern, applying appropriate controls without penalizing test takers for things that do not actually matter.

This is what separates a system that generates noise from one that produces clarity. Fewer false positives. Less administrative burden. More defensible outcomes.

Evidence That Holds Up

When a result is challenged, the question is not whether the system flagged something. It is whether a person reviewed it, made a judgment, and documented their reasoning.

Integrity Advocate's human review process ensures that every flagged session is assessed by a trained reviewer before any outcome is recorded. That is what makes the result defensible, and what allows programs to stand behind what they issue.

Built for What Comes Next

AI cheating tools are the most urgent current threat, and Integrity Advocate's Counter-AI capability addresses them directly. But a sustainable integrity strategy cannot be built around detecting one category of tool. It has to be designed to adapt as threats evolve.

That is the difference between a feature and a framework.

{{post-stat-highlight}}

The full white paper walks through exactly how this works in practice, including how each control layer functions, how Counter-AI fits within the broader system, and what proportional enforcement looks like for programs of different sizes and risk levels.

{{post-cta}}