AI Tools Struggle to Cheat Soft Skills Assessments to Identify Top Talent, New Cangrade Research Finds

WATERTOWN, Mass. – April 8, 2025 – Cangrade today announced the results of its first AI-Enabled Candidates in Hiring Report, exploring the performance of AI tools, like ChatGPT, “cheating” the interview process. The findings suggest that while job seekers are leveraging generative artificial intelligence (AI) to enhance their applications, AI tools struggled to successfully navigate Cangrade’s soft skills assessments designed to identify top talent.

AI is reshaping the job market, making it easier for both recruiters to market roles and for candidates to apply to jobs. This has resulted in a significant increase in the number of applicants per role, but a decrease in serious, qualified candidates. To stand out among the masses, people are turning to AI to draft resumes, generate interview responses, and complete skills assessments. 

Although it seems to be a smart solution to highlight keywords and metrics desirable to recruiters for a specific role, AI may actually be doing job seekers a disservice. In fact, Cangrade’s research showed that AI-assisted applicants fail to effectively perform on hiring assessments designed to measure job-relevant personality traits.

Other key findings include:

- AI-generated responses had only a 12% success rate in passing Cangrade’s hiring assessment, a figure comparable to acceptance rates at top universities.

- ChatGPT consistently avoided extreme responses, favoring cautious agreement over strong opinions. If a role that requires high attention to detail, like an accountant, “slightly agrees” with the statement, “I find it hard to focus,” this could negatively impact the candidates’ chances of moving forward. 

- AI prioritized traits like persistence, goal achievement, and prosocial behavior, while steering clear of “negative” interpersonal traits. This would be counterproductive for jobs focused on individual contributions vs. teams. 

- AI struggled with job-specific nuances, leading to misalignment with role requirements, reducing its effectiveness in passing assessments.

The study underscores the need for HR professionals to adapt their hiring processes to identify and mitigate AI-enabled application inflation. While banning AI outright is impractical, organizations can implement key strategies to ensure hiring accuracy, including:

Leveraging soft skills assessments that lack clear “right” or “wrong” answers.

Implementing AI-resistant hiring techniques, such as time-limited assessments, video interviews, and AI-detection tools.

Reducing reliance on resumes and cover letters, which are easily manipulated by AI.

“As AI becomes more embedded in the job application process, HR teams need to evolve their hiring strategies to distinguish genuine top talent from AI-enhanced applications,” said Gershon Goren, founder and CEO, Cangrade. “Our research shows that well-designed soft skills assessments remain an effective first line of defense against AI-enabled job seekers overstating their abilities.”

To access the full report, click here. For more information about Cangrade’s AI-powered, bias-free hiring and talent management solutions, visit www.cangrade.com.

About Cangrade

For HR leaders, Cangrade is the bias-free, AI-powered talent intelligence platform. By integrating data into talent acquisition and management processes, Cangrade enables businesses to make strategic and efficient decisions from initial screening through the entire employee lifecycle. Delivering 10x more accurate predictions of talent success and retention than traditional methods, the company’s Pre-Hire Assessment has helped organizations like Wayfair, FDNY, Lamar Advertising, and Applied Industrial Technologies make the right hiring decisions for over 10 million candidates and counting. For more information, visit www.cangrade.com.

Media Contact:

Gina Devine

Public Relations

[email protected]