As Test Takers Turn to AI for Answers, Exam Security Must Catch Up

Certifications play a critical role in keeping up with industry developments, but artificial intelligence is now muddying the verification process. Between Meta Glasses, AI browsers, and deepfakes, exam administrators face growing risks to monitor.

Score legitimacy has become harder to assess as test takers use AI to fill in the blanks. As a result, the industry is shifting toward more in-person testing and stricter fraud-detection protocols to keep certifications credible.

New Security Challenges

Cheating has always been a risk in testing, but AI has significantly changed the game for IT certification, explains Marielle Carlos, data forensics scientist at Caveon. Traditional threats like harvesting pre-knowledge, plagiarism, or collusion have evolved as people use AI as an unauthorized aid.

Carlos has observed a "huge boom of people" who are testing fast and receiving high scores. While there's no concrete data that it's all due to AI, the trend has raised enough organizational concern to warrant further investigation.

Stricter Protocols

Widespread irregular exam results and verified cheating attempts have led ISC2 to require in-person testing for all exams, according to Casey Marks, chief operating officer at ISC2. They've implemented "the most stringent security protocols available" including biometric and identity verification.

Marks notes the testing environment mirrors the most stringent conditions used by licensed professions like doctors. "The candidate, the employer, and every other professional holding that credential depend on the credibility of the process," he adds.

Detection Methods

Research by Sarah Toton, director of data forensics at Caveon, has revealed distinct differences in response timing between AI and human test takers. "When someone uses AI, they're often six to 48 times faster compared to an individual test taker," Carlos explains.

However, AI can get some things wrong, particularly questions involving human judgment. Carlos emphasizes the need for a layered approach to protecting score validity across the entire process, from secure exam design to result analysis.

Certifications Evolve

Despite the cheating risks, employers haven't shifted their view of certifications when making hiring decisions. Many certifications, like ISC2's CISSP, continue to evolve to address how AI impacts security best practices.

New AI-specific security certifications are entering the market, such as the Computing Technology Industry Association's SecAI+, which helps organizations secure, govern, and responsibly integrate AI into their cybersecurity operations. As certifications evolve, the value for employers is increasing, Marks adds.