November 7, 2025
|
5 min read
The 2025 Conference on Test Security brought together higher education institutions, certification organizations, legal experts, and technology vendors around a shared challenge: maintaining assessment integrity in an AI-driven world. Integrity Advocate CEO Brandon Smith shares his key takeaways, covering the growing non-negotiability of exam validity and security, the evolving legal landscape around misconduct and civil disputes, the uneven ethical rigor of AI tools in assessment, and the increasingly critical balance between test security and learner privacy.

The 2025 COTS Conference brought together a diverse mix of professionals from higher education, certification organizations, legal experts, and technology vendors all united around a shared challenge: maintaining the integrity of assessments in an increasingly digital and AI-driven world.
Across both higher education and professional certification sectors, one theme dominated the conversation; security and validity are non-negotiable. Institutions are doubling down on technologies and practices that ensure every test taker is who they say they are and that the results accurately reflect their own work. With remote and hybrid testing becoming the norm, maintaining that integrity has never been more critical, or more complex.
A recurring topic throughout the conference was the legal ramifications of exam misconduct. When test takers are caught cheating or engaging in fraudulent behavior, the response can’t be improvised. Organizations need clear, documented legal frameworks to protect themselves and their brands.
With the rise in reported cheating incidents comes an increase in interactions with test takers around alleged violations some of which escalate into accusations, libel, or civil disputes. The consensus: it’s essential to plan ahead, not react after the fact.
AI was everywhere at COTS. While some vendors and institutions are embracing AI in a responsible and transparent way, others are less rigorous about the ethical guardrails they apply. The takeaway? Not all AI tools are created equal.
Decision-makers must ask tough questions, vet vendors carefully, and understand how AI is being used especially when it impacts privacy, proctoring, and identity verification. Blind trust in “smart” systems can lead to unintended consequences.
Perhaps the most nuanced discussions centered on finding the right balance between test security and user privacy. With test takers increasingly sensitive to how their data is collected and used, organizations are being challenged to uphold both integrity and privacy, without compromise. It’s not an either/or decision; the future of assessment requires innovative solutions that respect both principles.
COTS continues to position itself as a true thought leadership event. Every session was rich with real-world examples, backed by data, and framed within professional and legal contexts. The mix of participants, from educators to legal professionals to technology providers, created a vibrant ecosystem of perspectives, debate, and shared problem-solving.
The future of testing isn’t just about better technology; it’s about smarter, more ethical implementation. The conversations at COTS made it clear: the organizations that thrive will be those that treat integrity as both a value and a strategy. As AI reshapes the landscape and legal accountability intensifies, preparation, transparency, and ethics will define the leaders in this space.
{{post-cta}}
Find answers to the most commonly asked questions from our clients.
The Conference on Test Security is an annual event that brings together professionals from higher education, professional certification, legal practice, and assessment technology to address the challenges of maintaining exam integrity in a digital and AI-driven environment. COTS is considered one of the leading thought leadership events in the assessment security space.
Five themes dominated the 2025 conference. Security and validity are non-negotiable, with institutions doubling down on technologies that confirm identity and ensure results reflect genuine knowledge. The legal landscape is changing, with misconduct incidents increasingly escalating into civil disputes that require documented legal frameworks rather than improvised responses. AI tools vary significantly in their ethical rigor, and decision-makers must vet vendors carefully. The balance between test security and learner privacy requires solutions that uphold both principles simultaneously. And the organizations that lead will be those that treat integrity as a strategic value rather than a compliance obligation.
As cheating incidents increase and remote testing expands, disputes around alleged violations are more frequently escalating into formal accusations, libel claims, and civil proceedings. Organizations that handle misconduct without clear, documented legal frameworks expose themselves to significant liability. The COTS 2025 consensus was that preparation before an incident occurs, including transparent policies and defensible documentation, is essential.
Not all AI tools apply the same ethical standards or transparency to how they detect, flag, and report behavior. Decision-makers should ask how AI is used in the flagging process, whether human review is applied before decisions are made, how bias and false positives are managed, what data is collected and for how long, and whether the vendor can clearly explain every decision the system produces. Blind trust in automated systems without understanding these details creates risk for both institutions and test-takers.
The most important insight from COTS 2025 on this topic is that security and privacy are not in conflict when a system is designed correctly from the start. Privacy-first proctoring collects only what is necessary, retains data only as long as required, and applies human judgment rather than broad surveillance to every decision. Organizations that treat this as an either/or choice are working with the wrong framework. The right approach is one that upholds both principles simultaneously through intentional design.