The hallways of the I.C.E. Exchange are usually buzzing with a frenetic kind of energy. It’s a blend of psychometric passion, policy debates, and the warm familiarity of colleagues reconnecting. This year in Phoenix, the buzz was still there, but it had a different frequency. There was a nervous undertone that permeated nearly every keynote, breakout session, and coffee-break conversation.
While the program promised deep dives into micro-credentials, global expansion, and operational efficiency, one theme overshadowed all others to become the defining takeaway of the conference: Security. Specifically, the growing threat posed to certification integrity by rapidly evolving Artificial Intelligence and emerging technologies.
For years, the credentialing community looked positively on ways to deliver exams faster and issue badges more smoothly. However, the industry has now turned a corner where technology and innovation have grown so fast, that there is a fear that the thing that was once coveted should now be feared.
Here is why security in the age of AI became the inescapable narrative of this year’s I.C.E. Exchange, and what it means for the future of credentialing.
Next Level Cheating
The most chilling realization shared by experts across multiple sessions is the sheer accessibility of sophisticated cheating tools.
In the past, organized exam fraud required effort, connections, and usually significant money. Today, finding effective ways to cheat is as simple as doing a web search. We heard terrifying anecdotes about the evolution of remote proctoring circumvention. It’s moved beyond simple screen-sharing hacks. We are now entering the era of real-time deepfakes used to bypass identity verification, and unobtrusive smart-wearables that can capture exam content and feed it to an AI, which then whispers the answer back to the candidate.
The speed at which these adversarial technologies are advancing is outpacing our traditional defense mechanisms. A security patch applied today might be obsolete by next month’s AI model update. This might be the biggest threat online proctoring has faced to date.
The Crisis of Validity
Why did this topic suck the air out of the room? Because if security goes, everything else goes with it.
The foundational currency of any certification or licensure body is trust. A credential tells employers, regulators, and the public that an individual possesses the knowledge and skills required to perform a job safely and competently.
If AI tools make it effortless for unqualified individuals to pass high-stakes exams, the validity of that credential evaporates. When validity dies, public protection is compromised. I.C.E. 2025 was a stark reminder that we are facing a potential industry shattering event for credential integrity.
Fighting Back
But, this industry is resilient. Instead of retreating, the conference focused on the ways to fight back. The most valuable sessions pivoted away from admiring the problem and toward actionable solutions.
The consensus among thought leaders was that trying to simply block AI is a losing battle. We cannot create a hermetically sealed digital environment anymore. Instead, the industry must evolve rapidly in two directions:
1. Rethinking Assessment Design: The era of the knowledge-recall multiple-choice question is drawing to a close. If ChatGPT can answer a test item correctly in three seconds, that item is no longer measuring human competence.
Sessions on psychometrics emphasized the urgent need to move toward performance-based testing, simulations, and complex situational judgement tests. In other words, formats where the answer isn’t a fact to be retrieved, but a process to be demonstrated. We need to assess skills that AI currently struggles to replicate authentically, such as nuanced professional judgment, ethics in gray areas, and tangible skills application.
2. Fighting Fire with Fire: The other side of the coin is using AI to save us from AI. Several vendors showcased emerging tools that use machine learning not to grade tests, but to secure them. This involves analyzing thousands of data points during a remote exam, including keystroke dynamics, eye movement patterns, and response latencies, to flag anomalies that human proctors would never catch. We need AI-driven forensics to detect the fingerprints of AI-driven cheating.
The New Normal
Leaving I.C.E. Exchange 2025, the overarching feeling was one of urgent clarity and optimism. Now more than ever, security is in the forefront and is a continuous, central business strategy. Attendees left with a newfound sense of hope. What once was doom and gloom is now a feeling towards a brighter future.
This is a defining moment in our industry, and those companies willing to rise to the challenge will succeed. At Kryterion, we are ready to take the new forms of AI cheating head on. Talk to us about how we can help your credentialing program retain its value.
About Kryterion
Kryterion Inc. is a global leader in innovative testing and credentialing solutions, helping organizations across various sectors develop and manage their assessments with our advanced test development platform and multi-modal delivery solutions. Established in 2001, Kryterion offers secure, integrated services and extensive support, empowering candidates to demonstrate skills and achieve success in world-class careers.