Our product and engineering teams are focused on researching the impact and exploring new solutions to generative AI tools like ChatGPT so our customers can continue to hire with confidence. As an immediate action, we have implemented a task tagging system based on our research-backed estimates of how solvable the task is by ChatGPT.
We have manually tested our entire Codility programming task library using a standardized testing protocol developed by our engineers. Each task in our library is now tagged based on our estimates of its AI solvability. Values for this filter can be High, Medium, or Low. Tasks with High resistance are more challenging for artificial intelligence, and solutions generated by AI prompts will usually fail or be vastly incomplete.
As our knowledge of AI prompt systems grows, we constantly monitor and revamp our library tasks' integrity. If you are concerned about unqualified candidates being able to solve your tests using those systems, pick tasks with High resistance.
All newly released Codility tasks are immediately and thoroughly tested against generative AI tools - we also recommend using the “New Tasks” filter together with the “AI Resistance” filter for best results.