Proctorio Blog

We recently attended the 2026 ATP Innovations in Testing Conference in early March, which brings together many of the organizations responsible for designing, delivering, and protecting professional and educational assessments.

One thing was immediately clear. People in the testing industry are worried. Across sessions, panels, and hallway conversations, the same issue kept coming up again and again.

How do we know that the person taking an exam is actually the one doing the work?

For decades, testing systems largely relied on a simple assumption. The person sitting for an exam was the one answering the questions. Today, that assumption is under pressure.

A Poll That Captured the Mood

During one session, attendees were asked a simple question. “To protect exam integrity in the AI era, should we assume every test taker could be using AI unless proven otherwise?” The results were revealing.

47% agreed
15% strongly agreed
21% disagreed
6% strongly disagreed
11% were neutral

That means 62% of the room leaned toward agreement.

The point was not that most people think candidates are cheating. The point was that the tools now exist that make it possible. And that has changed how people in the assessment industry think about the problem.

Cheating Continues to Grow into a Mature, Pseudo-Legitimate Market

One of the most surprising discussions at the conference was how organized cheating has become.

This is no longer just about someone glancing at their phone during a test. Entire businesses now exist that help people cheat on exams.

These services advertise openly online. Some offer real time help during exams. Others sell AI tools that generate answers instantly. Some even offer proxy testing where someone else takes the exam for you.

One example shared during a session showed a service offering to take the LSAT exam on a student’s behalf for $10,000. The package included technical setup, coaching, troubleshooting during the exam, and guidance on how to avoid being detected.

In other words, cheating has moved from being a student problem to being a business model.

The Tools Are Getting Cheaper and Easier

Another theme that came up repeatedly was how accessible cheating technology has become.

Some of the tools are surprisingly simple. For example, presenters demonstrated how an HDMI splitter can mirror a test taker’s screen to another computer in a different room. Someone watching remotely can then signal the correct answers. These devices cost around $30 and are sold at major retailers.

Other tools are more advanced. Smart glasses can stream video to another person. AI systems can listen to questions and suggest answers instantly. Browser extensions can quietly connect to large language models during a test.

And because these tools are easy to build, new versions appear constantly.

One speaker made the point that a technically savvy student can create a new cheating app in a single afternoon by “vibe coding.” By the time testing platforms block one tool, several new ones may already exist.

The Data Shows How Sophisticated Cheating Can Be

Testing organizations are increasingly relying on data to detect suspicious behavior.

In one example shared at ATP, investigators discovered that more than 500 stolen test questions had been circulating through a paid subscription service online.

In another case, analysts identified groups of candidates whose answer patterns were almost identical. The probability that those responses occurred naturally was described as less likely than winning the lottery twice in a row.

Another organization reported detecting more than 60,000 suspicious testing events across different exam programs using automated analysis tools.

These kinds of signals can include extremely fast completion times, identical wrong answers across candidates, sudden jumps in performance, or unusual changes in answers after breaks.

None of these proves cheating on their own. But together they can reveal patterns that are difficult to explain otherwise.

Remote Testing Still Matters

Despite all these challenges, one point came up again and again. Remote testing is not going away and shouldn’t.

It allows people to take exams from anywhere. It expands access to education and professional certification. It reduces travel costs and scheduling barriers.

For many organizations, remote testing has made assessments possible at a much larger scale. The challenge is not whether remote testing should exist. The challenge is how to ensure the results remain trustworthy.

The Tests Themselves May Need to Change

Another interesting theme from ATP was that security may not come only from monitoring. It may also come from changing the design of exams themselves.

Some testing programs are experimenting with assessments where every candidate receives a different set of questions. Others are exploring simulations, conversations, or scenario based questions instead of simple multiple choice tests.

One example discussed at the conference was a language exam where no two candidates receive the same test. While this approach isn’t new, questions are generated dynamically, making it nearly impossible to reuse stolen answers. If every test is unique, the incentive to harvest questions drops dramatically.

No One Can Solve This Alone

A clear takeaway from the conference is that this problem is bigger than any single company or technology. Testing organizations, universities, technology platforms, and policymakers all have a role to play.

Some industry leaders are now pushing for laws that make it illegal to advertise or sell cheating services online, similar to legislation that already exists in countries like the United Kingdom and Australia.

But regulation alone will not solve the problem.

The industry will need new technology, new assessment designs, and stronger collaboration across organizations.

The Question Behind Everything

In the end, the conversation at ATP kept returning to a simple question.

For centuries, education and professional credentialing relied on trust. If someone passed an exam or earned a certification, it meant they had demonstrated the knowledge themselves.

But in a world of AI tools, remote work, and digital testing, that assumption can no longer be taken for granted. The real question the industry is now trying to answer is this.

How do we verify that the work we see is actually the work of the person who submitted it? That question will shape the future of testing, education, and professional certification for years to come.

The good news is that Proctorio has seen this coming for a while, and will be launching new offerings to attack this as well as other sticky obstacles in 2026. Stay tuned.