Generative AI has made it easy to improve a résumé, draft a cover letter, prepare for an interview, and produce polished answers. Those uses are not automatically improper. The hiring challenge is that application materials can become less informative about what a person can do independently.
Gartner reported in 2025 that 39% of surveyed candidates used AI during the application process. Among those users, some reported using it for assessment answers. At the same time, only 26% of surveyed candidates trusted AI to evaluate them fairly. That tells me we have two problems to solve at once: employers need trustworthy evidence, and candidates need a process they can trust.
Integrity should protect evidence—not create automatic suspicion
A useful integrity program distinguishes between a review signal and a conclusion. Leaving an assessment window, triggering a candidness indicator, or creating a pattern that requires closer review does not by itself prove cheating or dishonesty.
The purpose of an integrity indicator is to improve context. It can help an authorized reviewer decide whether the result is interpretable, whether follow-up is appropriate, or whether the participant should be given a chance to explain what happened.
Layer safeguards according to the role and risk
Not every assessment situation needs the same level of verification. A developmental conversation and a high-security, safety-sensitive selection process do not carry the same consequences. The safeguards should be proportional to the intended use, the role, the testing environment, and the organization's policy.
- Restrict cut-and-paste activity to help protect assessment content.
- Record outside-window activity as review context rather than an automatic failure.
- Use candidness or response-pattern indicators to guide cautious interpretation.
- Add identity verification when the situation justifies the additional step.
Candidate trust improves when the process is understandable
Organizations should explain what the assessment is for, what candidates can expect, and how information will be used. When additional integrity controls are present, the process should avoid surprises and provide a path for reasonable questions or accommodations.
The strongest response to AI-assisted applications is not simply more surveillance. It is better evidence: job-relevant assessment content, consistent administration, proportionate safeguards, structured interviews, and human review that considers the complete context. The objective is confidence, not suspicion.
Practical takeaways
What talent leaders can do next
- Treat integrity indicators as review signals, not verdicts.
- Match safeguards to the consequences and context of the assessment.
- Give candidates clear expectations and an appropriate path for questions.
- Combine integrity information with job-related evidence and structured human follow-up.
Sources and related guidance
- Gartner: Candidate trust, AI use, and application integrity
- ODNA Talent: Assessment Integrity
- ODNA Talent: Online Assessment Integrity Guide
Educational information only. Organizations remain responsible for obtaining appropriate legal, scientific, privacy, accessibility, and professional guidance.