AI in testing

Artificial Intelligence in testing

Transforming exams and assessments

AI innovation with integrity

AI in testing is transforming how exams and assessments are developed, secured, and delivered. At PSI, we apply AI to strengthen test credibility, fairness, and efficiency across the entire assessment lifecycle.

AI has the potential to accelerate test development, strengthen exam security, enhance candidate experience, and optimize program operations. But its use must be transparent, responsible, and built on best practices. That’s where PSI stands apart.

With decades of experience, global reach, and deep expertise in licensure and certification testing, PSI is pioneering the thoughtful application of AI in assessment.

On this page you will learn:

Why AI in testing matters

AI is no longer just a buzzword in testing. At PSI, we are already applying AI to:

Detect sophisticated test security threats such as deepfakes and identity fraud.
Accelerate content generation and scale item banks.
Efficiently develop practice tests and test preparation materials.
Deliver operational insights that improve program outcomes.

For credentialing bodies, licensing boards, and employers, AI delivers not only efficiency but also scalability, sustainability, and strengthened trust in a rapidly changing testing environment.

PSI’s approach to AI

At PSI, AI is not a bolt-on. It is embedded holistically across the assessment lifecycle with human oversight at its core.

Collaborative and Human-in-the-Loop (HITL): AI supports Subject Matter Experts (SMEs), psychometricians, and proctors. It augments rather than replaces human expertise.
Ethical and transparent: We align every application of AI with industry best practice, fairness, and compliance.
Built for assessment: Unlike generic tools, PSI’s AI is designed to align with testing and measurement principles.
Secure and responsible: Data is protected within closed systems and governed by strict security protocols. Clients retain ownership of AI training data and test content generated.

Now, as part of the ETS family, PSI combines agility and innovation with the scale and authority of a global leader in education and assessment.

Applications of AI in testing across the assessment lifecycle

AI-enhanced test security

Protecting the integrity of exams is paramount. AI enhances PSI’s already comprehensive test security by:

Powering AI-enabled online proctoring.
Detecting digital document forgeries.
Identifying deepfake attempts.
Analyzing patterns to uncover potential fraud, such as organized cheating rings.

Our AI-enabled systems work with trained human proctors, data forensics indices, and web monitoring to stay ahead of emerging threats. This layered approach ensures exam integrity in both remote and test center environments.

244% rise in digital document forgeries in 2024.

Deepfake attempts occurred every five minutes globally in 2024.

2025 Identity Fraud Report, Entrust Cybersecurity Institute

AI-driven test development

AI is revolutionizing the way assessments are built. With PSI’s approach:

AI accelerates item generation, creating valid and defensible content at scale.
SMEs are empowered as ‘content architects’, focusing on oversight, quality assurance, and psychometric soundness.
Human-in-the-Loop (HITL) processes ensure every item meets rigorous standards of fairness, validity, and relevance.
Diverse item variants strengthen test security and expand item banks cost-effectively.
Both the data that informs AI and the content created are protected in a closed, secure system.

The result? More efficient development, secure assessments, and tests designed to capture harder-to-measure skills such as critical judgment and interpersonal effectiveness.

Proven performance in AI test development

Comparable quality
77.4% of AI-generated test items met all psychometric thresholds compared to 75.5% human-authored.

High SME confidence
65.6% of AI-generated test items approved for pretesting, SME item retention improved by 10% between batches.

Time and cost savings
Faster throughput with no compromise on rigor.

AI in the test taker experience

Candidates are at the heart of everything we do at PSI. We harness AI to create more equitable and supportive testing journeys:

AI-powered test preparation: Tailored apps and AI learning assistants (AILA) that help candidates prepare with confidence.
Personalized guidance: Chatbots and self-service tools provide timely, accurate support.

By combining testing best practice with intelligent tools, PSI ensures candidates experience efficiency without compromising fairness and accessibility.

AI in program operations and analytics

For testing programs, AI offers actionable insights:

Predictive analytics that surface trends in candidate performance and emerging security risks.
Automated reporting to speed up decision-making.
Fraud trend detection to strengthen long-term program resilience.

These tools allow credentialing bodies to manage programs more efficiently while maintaining transparency and trust.

Why work with PSI on AI?

Choosing PSI means partnering with a leader that brings:

Proven scale: Over 28 million assessments delivered annually, across 400+ programs in 150+ countries.
ETS partnership: Backed by investment, scale, governance, and research leadership.
Deep expertise: 75+ psychometricians and test developers dedicated to scientific rigor, 70+ test security experts across PSI and ETS.
Holistic services: End-to-end assessment solutions or modular services tailored to your needs.
Innovation and integrity: The ability to innovate quickly without compromising test fairness or public safety.

FAQs about AI in testing and assessment

At PSI, we take a transparent, human-centered approach to AI in testing, ensuring innovation always goes hand in hand with psychometric rigor, fairness, and security. Below, we’ve answered some of the most common questions from credentialing bodies, licensing boards, and program leaders.

How does PSI use AI in test development for exams and testing programs?

PSI uses AI to assist in test item generation, allowing SMEs to focus on oversight and refinement. This AI in testing approach accelerates development, increases item variants, and ensures validity and fairness.

Does AI replace SMEs in AI-driven test development?

No. AI enhances test development efficiency, but SMEs are still essential as ‘content architects’ guiding, verifying, and refining content.

How does PSI protect data when AI is used in test development?

PSI’s AI test development operates within a closed, highly secure environment. All information used to train the system and all content it generates is isolated by client and fully protected from external access. This ensures security, privacy, and trust throughout the process.

How does PSI detect deepfakes and AI-assisted fraud in test delivery and exam security?

Our test security systems integrate AI-driven detection of digital forgeries, biometric manipulation, and synthetic IDs. Deepfake attempts are flagged by AI for further human review.

Is AI in test proctoring and exam monitoring fair and unbiased?

Yes. When AI in test proctoring augments – not replaces – human oversight, it ensures fairness and defensibility. Suspicious behaviors are flagged by AI for human review, maintaining transparency and candidate trust.

How does AI improve the candidate experience in tests, exams and assessments?

By providing personalized test preparation, AI-powered guidance, and personalized learning assistants, candidates benefit from more relevant and supportive tools.

What role does human oversight play in PSI’s applications of AI in testing?

Human judgment is built into every step of the assessment journey, from content review to proctoring, ensuring testing and ethical standards are never compromised.

How does PSI ensure compliance and regulatory alignment?

PSI aligns AI use in testing with global standards and regulatory frameworks, combining legal defensibility with operational security. PSI conforms with relevant global standards such as ISO, GDPR, and industry-specific regulatory frameworks.

How does PSI ensure AI in testing is explainable and transparent?

PSI designs every AI process to be fully auditable. Each AI-assisted output is linked back to source materials and documented through clear audit trails. This ensures SMEs and program leaders can always see how an item or decision was generated, making the system defensible, transparent, and aligned with industry and regulatory expectations.

What are the limits of AI in assessment?

AI is a powerful tool, but it is not a substitute for human judgment or psychometric expertise. At PSI, we emphasize a Human-in-the-Loop (HITL) approach, where SMEs, psychometricians, and test security experts guide and oversee every step. AI accelerates processes and expands scalability, but final decisions about validity, fairness, and defensibility always remain with human experts.

Can PSI’s AI solutions be tailored to my industry or jurisdiction?

Yes. PSI’s AI testing solutions are adaptable across industries such as healthcare, real estate, finance, and IT, and configurable to jurisdiction-specific requirements. Whether global certification, state or national licensure, solutions can be tailored to meet your regulatory environment.

Partner with PSI on AI

Artificial intelligence is transforming assessment. With PSI, you gain a partner who balances innovation with integrity, ensuring that AI strengthens the credibility of your program