Technical hiring platform Codility announced a new library of content that mitigates cheating on skills assessments by using ChatGPT. The product will help companies create tests that are ChatGPT-resistant or identify and filter out ChatGPT-solvable tasks. The company said it is now the only leading assessment vendor actively enabling customers to create AI-resistant assessments.
Founded in 2009, Codility helps organizations test job seekers’ technical coding skills and make evidence-based hiring decisions. The company said it has worked with OpenAI since 2020 to get insights on generative AI and its relationship to coding assessments.
The company found that the recent rise of widely available AI technology — like ChatGPT — has presented new challenges for hiring. It follows, then, that employers and providers must adapt to the changing needs of the industry.
“From allowing engineers to showcase their skills using generative AI to introducing features that mitigate and detect its use, we will continue to provide our customers with more options to select the assessment methods that best align with their hiring philosophy and goals, supporting building diverse and high performing engineering teams,” said Codility CEO Natalia Panowicz.
Addressing Changes in Tech
The new offering tags tasks according to their potential solvability by ChatGPT. That allows organizations to create coding tests that more accurately assess the skills of prospective employees.
In addition, Codility uses its existing database of solutions — submitted by candidates and testers — to check assessments and determine if a solution was previously submitted. The company has also evolved how it develops new content and is increasing the production of highly ChatGPT-resistant task types within the assessment platform. According to the company, creating fresh content for its assessments on a regular basis can also prevent the possibility of candidates cheating.
“The ongoing testing and evolution of our AI-resistant library will support hiring teams in making confident, evidence-based hiring decisions, even with the growing prevalence and use of AI in skills testing,” said Panowicz.