Artificial intelligence can now answer technical questions, generate reports, interpret data, write software, analyze images and assist professionals in making increasingly complex decisions. That creates a fundamental challenge for personnel certification: when a candidate performs with the assistance of AI, whose competence are we actually assessing?
For decades, personnel certification has been built around a relatively straightforward principle. A certification body defines the competence required for a profession or activity, assesses the individual against those requirements and, when the requirements are fulfilled, issues a certificate confirming competence.
Artificial intelligence complicates that model.
The professional of the future may not perform every task independently. Engineers, inspectors, auditors, cybersecurity specialists, analysts and other professionals may increasingly work with AI systems capable of performing parts of the technical process themselves.
Personnel certification therefore faces a new question: how should competence be demonstrated when human capability and machine capability increasingly operate together?
Personnel Certification Has Always Been About Confidence
The purpose of personnel certification is not simply to issue credentials.
It is to provide confidence that a certified individual fulfils defined competence requirements within a specific certification scheme.
Employers, regulators, clients and asset owners rely upon that confidence when certified personnel perform activities capable of affecting quality, safety, compliance, reliability or professional outcomes.
If artificial intelligence performs an increasing portion of a professional task, certification schemes need to determine which capabilities must remain demonstrably human and which capabilities may legitimately be performed with technological assistance.
ISO/IEC 17024:2026 Brings AI Directly Into the Conversation
The publication of ISO/IEC 17024:2026 is particularly significant because the revised international standard explicitly addresses the use of artificial intelligence within personnel-certification processes.
The standard includes dedicated requirements concerning AI and establishes expectations around monitoring and validating AI-generated outcomes, maintaining appropriate human oversight and ensuring competence in the use of AI tools.
This reflects a broader reality: artificial intelligence is no longer external to conformity assessment.
It can now participate in activities associated with examination, assessment, administration, monitoring and potentially decision-support processes within certification systems.
The individual performs the relevant task and the certification process evaluates whether that individual possesses the required knowledge, skills and abilities.
Technology may contribute to analysis, drafting, calculation, diagnosis or decision support, creating a need to distinguish human competence from machine-assisted performance.
The First Challenge: AI Can Produce the Correct Answer
Traditional examinations often assume that a correct answer provides evidence of candidate knowledge.
Generative AI challenges that assumption.
A candidate may be able to obtain technically sophisticated answers, calculations, interpretations or written explanations without independently possessing the competence that the examination intends to measure.
This creates an important distinction between producing a correct outcome and demonstrating competence.
Knowledge Recall May Become Less Important
Many traditional certification examinations place significant emphasis on recalling technical information.
But professionals increasingly operate in environments where standards, regulations, databases, digital tools and AI assistants are continuously accessible.
This raises an important question for scheme owners:
Should certification continue to measure how much information a professional can remember — or increasingly measure whether that professional can interpret information, verify it, challenge it and apply it correctly?
In many professions, competence may increasingly shift toward higher-order capabilities.
AI Literacy May Become Part of Professional Competence
Preventing professionals from using artificial intelligence may not always reflect the workplace in which they actually operate.
In many sectors, competent professionals may increasingly be expected to use AI effectively.
The certification question then changes.
Instead of asking whether a candidate can perform a task without AI, a certification scheme may sometimes need to determine whether the candidate can use AI responsibly while remaining professionally accountable for the outcome.
Recognize that AI-generated outputs can be inaccurate, incomplete, biased, outdated or inappropriate to the specific technical context.
Independently evaluate AI-generated information before relying upon it in professional work.
Understand when AI assistance is suitable and when professional analysis must remain independent.
Recognize confidentiality, privacy, intellectual-property and cybersecurity implications when using external AI systems.
Avoid inappropriate automation bias and remain capable of challenging machine-generated recommendations.
Understand that use of an AI tool does not automatically transfer professional responsibility to the technology.
Should AI Be Allowed During Certification Examinations?
There may no longer be one universal answer.
The correct approach should depend upon what the certification scheme intends to measure.
AI access may be prohibited where the objective is to demonstrate that the candidate personally possesses essential knowledge, calculation ability, interpretation skills or safety-critical judgement.
AI may be permitted where responsible use of digital tools forms part of the real professional activity and the assessment is designed to evaluate verification, judgement and accountability.
A sophisticated certification scheme may ultimately require both.
Candidates could first demonstrate essential independent competence and then demonstrate their ability to perform professional tasks using authorized AI tools under controlled conditions.
Practical Assessment Becomes More Valuable
AI increases the importance of assessment methods capable of examining how a candidate actually approaches a technical problem.
Multiple-choice questions remain useful in appropriate circumstances, but they may provide limited evidence of professional judgement where advanced AI tools can generate answers almost instantly.
Present realistic professional situations requiring candidates to evaluate incomplete or conflicting information.
Observe candidates performing technical or professional tasks rather than relying exclusively on knowledge questions.
Ask candidates to explain reasoning, assumptions, alternatives and limitations behind their conclusions.
Require candidates to distinguish credible evidence from weak, irrelevant or potentially incorrect information.
Provide imperfect AI-generated outputs and assess whether candidates can identify errors, omissions or unsafe recommendations.
Require candidates to justify why a technical or professional decision is appropriate and defend it under questioning.
The Competent Professional Must Be Able to Challenge the Machine
One of the most important future competence requirements may be the ability to recognize when an AI system is wrong.
Artificial intelligence can present incorrect information confidently. In highly specialized technical fields, the output may appear convincing even when important assumptions, standards, limitations or risks have been misunderstood.
Automation Bias Creates a New Competence Risk
Professionals can become overly dependent on automated recommendations simply because the output appears analytical or authoritative.
This phenomenon becomes particularly significant where AI is used in safety, inspection, engineering, auditing, cybersecurity, healthcare or other activities involving consequential professional judgement.
Personnel-certification schemes may therefore need to assess whether candidates understand:
Professional Responsibility Does Not Disappear
Artificial intelligence may produce analysis, recommendations or draft conclusions, but professional responsibility remains an important consideration.
In many regulated or safety-sensitive activities, an individual must ultimately decide whether information is sufficiently reliable to support an action or decision.
- Can the candidate explain how an AI-generated conclusion was reached?
- Can the candidate recognize when the output conflicts with technical requirements?
- Can the candidate independently verify critical information?
- Does the candidate understand when AI should not be used?
- Can the candidate identify uncertainty and escalate appropriately?
- Does the candidate understand confidentiality and data-security implications?
- Can the candidate remain accountable for the final professional judgement?
AI Can Also Enter the Certification Process Itself
Artificial intelligence is not only changing candidate behaviour.
Certification bodies themselves may increasingly use AI within certification activities.
Potential applications may include administrative processing, examination development support, remote monitoring, assessment assistance, anomaly detection, candidate interaction and analysis of certification information.
ISO/IEC 17024:2026 directly recognizes this development by establishing requirements concerning AI within certification processes.
Does the assessment still demonstrate what the individual personally knows, understands and can responsibly perform?
Are AI-generated outcomes appropriately validated, supervised and controlled so that certification remains fair and reliable?
Human Oversight Becomes Essential on Both Sides
AI therefore creates two different human-oversight requirements within personnel certification.
The candidate may need to demonstrate responsible oversight of AI used in professional work.
At the same time, the certification body must maintain appropriate oversight of AI used within its own certification process.
Examination Security Is Changing
Generative AI also creates practical examination-security challenges.
Traditional controls designed to prevent candidates from consulting books, websites or another person may not be sufficient when powerful AI assistance can be accessed through multiple devices and applications.
Certification bodies may therefore need to reconsider examination design rather than relying exclusively on stronger surveillance.
Establish confidence that the registered candidate is the individual completing the assessment.
Reduce dependence on questions that AI systems can answer without demonstrating genuine competence.
Review whether existing monitoring approaches remain appropriate for modern digital assessment environments.
Protect examination content and monitor potential compromise of assessment materials.
Establish appropriate processes for investigating unusual examination patterns without assuming that automated detection is conclusive.
Maintain competent human judgement where automated tools contribute to examination-security decisions.
Certification Schemes May Need to Be Redesigned
Existing certification schemes were often developed before generative AI became widely available.
Scheme owners should therefore consider whether existing competence criteria and assessment methodologies still represent the professional activity being certified.
The review should begin with the occupation itself.
- Which professional tasks are now routinely supported by AI?
- Which capabilities must still be demonstrated without technological assistance?
- Should competent use of AI become an explicit competence requirement?
- Which AI-related risks are relevant to the profession?
- Do current examinations measure recall when they should measure judgement?
- Should practical, oral or scenario-based assessment increase?
- How should certification address rapidly changing technology?
- Should recertification requirements include emerging digital competence?
Recertification May Become More Important
Artificial intelligence is developing far faster than traditional professional qualification cycles.
A competence profile that was appropriate several years ago may no longer adequately represent the tools, risks or responsibilities associated with the profession.
This may increase the importance of recertification and continuing competence.
Different Professions Will Need Different Answers
There should not be one universal AI competence requirement for every occupation.
An NDT technician, management-system auditor, cybersecurity professional, welding specialist, inspector, engineer and financial analyst interact with AI differently.
Certification schemes therefore need to remain profession-specific.
A broad awareness statement provides limited evidence that the person can use AI appropriately within a specific professional context.
The scheme defines how AI affects the particular occupation, associated risks, required human judgement and appropriate verification responsibilities.
Personnel Certification Could Become More Valuable — Not Less
It might appear that artificial intelligence reduces the importance of professional certification because machines can increasingly perform expert tasks.
The opposite may occur.
As AI becomes more capable, organizations may place greater value on independent evidence that the human supervising, interpreting and applying machine-generated information possesses genuine competence.
Certification can provide assurance that a person is not simply capable of operating an AI tool, but is capable of recognizing when its output should — and should not — be trusted.
What Certification Bodies and Scheme Owners Should Review Now
Organizations operating personnel-certification programmes should begin evaluating how artificial intelligence affects both the professions they certify and their own certification processes.
Determine whether current competence criteria still reflect how the profession is actually performed.
Identify assessment methods that may test information retrieval rather than genuine professional competence.
Establish clearly whether AI tools are prohibited, permitted or intentionally incorporated into particular assessments.
Determine whether verification, critical thinking, AI literacy or human-oversight capability should become part of the scheme.
Identify where artificial intelligence is used within certification activities and establish appropriate validation and oversight controls.
Ensure assessors and examination personnel understand how AI can affect candidate performance and assessment validity.
Reconsider examination security, remote-assessment controls and candidate-authentication arrangements.
Determine whether continuing competence arrangements adequately address technological change.
From Knowledge Certification to Competence Assurance
Artificial intelligence exposes an important distinction that has always existed in personnel certification.
Knowledge is not the same as competence.
A professional can possess information without knowing how to apply it. Conversely, a professional can access information through technology while still demonstrating exceptional judgement, interpretation and practical capability.
The challenge for modern certification schemes is therefore to determine which combination of knowledge, practical skill, professional judgement and technology-enabled capability represents genuine competence.
Who Are We Really Certifying?
That may become one of the defining questions for conformity assessment over the next decade.
Certification bodies will need to distinguish human competence from automated capability while recognizing that responsible use of technology may itself become part of professional competence.
The strongest certification schemes will therefore not attempt simply to exclude artificial intelligence from the assessment environment.
They will determine where independent human competence remains essential, where AI-assisted capability is legitimate and how both can be assessed without compromising confidence in the certificate.
The certificate of the future should not merely demonstrate that a person can work without artificial intelligence. It should demonstrate that the person remains competent, accountable and capable of professional judgement when artificial intelligence is part of the work.