Anti-counterfeiting measures in the age of AI
Strategies for meaningful aptitude diagnostics
AI is currently changing almost all areas of the working world and has long since reached the application and selection process in the context of aptitude diagnostics.
Particularly in the field of aptitude diagnostics, the question increasingly arises as to what impact these developments have on the validity and predictive power of selection procedures. Applicants can strategically use AI to optimize their self-presentation during the application process or to influence individual selection steps.
This results in various possible applications:
- AI can, for example, act as a digital assistant to support online tests or specialist tasks.
- Answers to interview questions can be pre-formulated or practiced.
- In extreme cases, even deepfake-supported video interviews would be conceivable.
Against this backdrop, the question of how tamper-proof aptitude assessment procedures actually are is increasingly being discussed in practice. And is the use of such procedures even worthwhile if "cheating" is theoretically possible?
Why good aptitude diagnostics are not replaced by AI, but become more important?
Cognitive aptitude tests remain among the most valid predictors of professional success and are therefore one of the most informative tools in personnel selection. Foregoing them would be risky, as poor hiring decisions quickly lead to high costs, process delays, and increased employee turnover.
At the same time, it's also true that other processes, such as interviews or traditional application documents, are increasingly influenced by AI. This makes it more difficult to reliably identify genuine skills and abilities.
And if we're being completely honest, anyone who really wanted to cheat could do so even before AI, for example by working on the procedures together with friends in front of the computer.
Therefore, good protective measures are crucial to ensure that your online test results remain reliable.
Anti-counterfeiting measures at INFO GmbH
To ensure protection against counterfeiting, we employ appropriate technical, organizational, and methodological measures in our online tests:
Test design
Even during test development, care is taken to ensure that tasks are not easily searchable, thus preventing direct access to solutions via search engines or AI. Additionally, task pools with comparable difficulty levels are used, from which individual tests are randomly generated. This prevents identical test versions and significantly hinders the pre-learning of individual items.
Technology
Access to the tests is granted via personalized and time-limited TANs, preventing unauthorized sharing. Additionally, the individual modules are time-limited, making it difficult to use external tools while completing them. If the test window is exited and another window is opened, for example, for using AI, applicants immediately receive a warning (mouse tracking).
Process / Communication
Even before the test begins, participants are transparently informed about the independent nature of the assessment and submit a corresponding declaration explicitly prohibiting the use of any aids. Furthermore, they are informed that a supplementary on-site test may be conducted during the subsequent selection process to verify the results.
Retest (anti-counterfeiting test)
For added assurance, a retest can be conducted within the company. This typically includes selected cognitive modules from the original test, particularly those related to abstract-logical reasoning. Since these abilities are considered relatively stable, they allow for a reliable verification of the test results. Significant discrepancies between the online test and the retest can indicate potential cheating.
Candidate experience and secure process design in the age of AI
Despite all the technological possibilities, transparency remains a key success factor in the selection process. Applicants should be able to understand at all times why certain measures are used and what expectations apply to the process. Clear instructions, such as those regarding the independent completion of online tests, the non-use of AI support, or the possibility of a retest within the company, create transparency and increase the acceptance of diagnostic procedures. In addition, declarations of independence and technical information strengthen commitment and ensure that expectations are clearly communicated from the outset.
At the same time, the process design should always be considered in relation to the specific applicant pool. With a large number of applicants, multi-stage selection processes with preliminary online assessments offer a particularly efficient method of pre-selection. This is precisely where online tests can demonstrate their strengths: those who cannot meet the requirements of an online test even with available resources will generally not be able to do so under controlled conditions either. For added assurance, a follow-up test on-site is recommended.
However, with a small number of applicants, the importance of each individual selection decision increases significantly. In such cases, it can be beneficial to integrate diagnostic elements directly into personalized formats, for example, during a recruiting day or an assessment day at the company. This allows diagnostic findings to be combined with personal impressions, realistic insights into the job, and measures to foster candidate retention.
The safeguards implemented to ensure test integrity should always be tailored to the specific procedure and process. Personality tests, for example, should be further validated and critically evaluated in an interview. The highest level of security regarding independent completion remains offered by procedures conducted under controlled, on-site conditions, as even proctoring-based solutions cannot guarantee complete control in an online context.
The increasing capabilities of AI do not mean, however, that aptitude diagnostics will lose importance or should be dispensed with entirely. Rather, current developments require a rethink in the design of selection processes: away from the question of whether diagnostic procedures should be used, and towards the question of how these can be meaningfully, transparently, and securely integrated into modern recruiting processes.
We would be happy to support you in developing suitable solutions for your individual situation – from creating a positive candidate experience and selecting appropriate diagnostic methods to developing secure and practical recruiting processes. Please feel free to contact us!
Image: AI generated with ChatGPT