As AI tools become increasingly prominent and accessible, maintaining the integrity of professional assessments is a top priority. To ensure fair and accurate results, our platform employs several multi-layered strategies designed to mitigate the use of AI and automated bots.
Strategic Question Design
The way questions are constructed is the first line of defense against AI assistance.
Avoiding Word-Based Vulnerabilities: Because AI models are built on language, “wordy” or purely text-based questions are easier for them to solve.
We prioritize number-based questions and graphs. AI often struggles to interpret these accurately and may “hallucinate” or provide incorrect answers that are unsupported by the data.
Curated Timers
Time management is a critical tool for deterring cheating.
Restricted Windows for Completion: Each assessment uses curated timers that are specifically calibrated to be quick.
This time constraint makes it difficult for a user to feed every question into an AI tool, as the process of capturing and entering data is often too slow to meet the deadline.
Image-Based Assessments
Integrating visual elements provides a significant hurdle for current AI technology.
AI Interpretation Errors: AI often struggles with specific visual details (sometimes producing errors like “six fingers” in generated images).
By using image-based questions, we create scenarios that are difficult for AI to interpret or solve correctly compared to standard text.
Technical Platform Restrictions
To prevent the rapid transfer of data to external tools, we have implemented strict technical safeguards.
No Copy and Paste: The assessment platform disables the copy-and-paste function.
This restriction prevents users from quickly moving assessment text into AI prompts or other websites.ard text.
Bot Verification
To ensure that a real person is taking the test, we include a verification step at the start of every session.
Bot Detection: Before the assessment begins, users must check a verification box to confirm they are not a bot.
Because automated bots cannot perform this specific manual interaction, they are effectively blocked from accessing the assessment. If our system detects a potential bot, further verification is required before allowing advancement to the assessment.