Candidates: Are you interviewing and need support?
It’s clear that ChatGPT is more than just a viral sensation soaking up its 15 seconds of fame. It’s here to stay. The AI tool has taken over the internet with images and copy-generation—and now it’s finding its way to the hiring process.
Many candidates are even using the tool to rewrite or optimize their resumes. And while some may consider this just a smart tool for candidates to add to their toolkit, at HireVue, we don’t believe resumes are the future of hiring—because decisions backed primarily by resumes are frequently decisions backed by unconscious bias.
But ChatGPT and hiring are blending beyond just resume creation. After reporting similar answers, recruiters are finding candidates are “cheating” on their interviews, whether it’s copying code generated by ChatGPT into a technical assessment or repeating the answers to OnDemand questions that they’ve fed to the AI chatbot.
And the issue won’t disappear. As ChatGPT continually becomes more sophisticated, hiring teams may need to alter some of their processes for better outcomes. We sat down with Scott Keller, Director of Solutions Engineering at HireVue, who gave 3 tips on how to combat cheating job candidates.
According to Keller, simply asking candidates how they arrived at their answers can provide valuable insight into whether a candidate came up with their own answers. For teams using HireVue’s technical assessments and OnDemand interviews, dive into the reasoning behind answers.
For example, hiring teams trying to avoid candidates who have simply copied and pasted code into a technical assessment, ask them how they generated their code in a live interview setting.
Keller notes that code generated by ChatGPT is getting more sophisticated. “ChatGPT has generated some interesting code in the past that has been flawed. However, we are now getting markedly stronger results that pass the vast majority of coding challenges. While there is typically a small amount of manual adjusting, it’s trivial.”
But there’s still some signs you can look for. For recruiters looking for any signs of cheating on code generation, the following would be a likely scenario of cheating:
For non-technical hiring, it can be a little harder to detect cheating, so you may start to notice similar patterns among candidate answers—and as ChatGPT becomes more sophisticated, it may be a little harder to decipher. So be sure to follow up in a live interview about how they arrived at their answers.
HireVue now offers Builder for Live which combines our structured interview generator with our live interview platform to ensure consistency and fairness in all virtual interviews. Builder for Live provides hiring teams with well-designed, pre-written, and science-backed interview questions that evaluate for job-relevant skills and assesses for competencies (not resumes).
Keller drives home what may be the most important concept around this topic.
“Hiring decisions should never be based on a single data point.”
At HireVue, we believe the best hiring decisions are made when humans and AI work together. AI is not a replacement for human decision-making, and no decision should be made on a single technical assessment or OnDemand interview.
Look at more than one data point to make an educated decision. From OnDemand interviews to various types of assessments to live interviews, HireVue offers solutions that ensure not only are all candidates evaluated the same, but empower TA teams to prioritize the right candidates.
Keller reminds us that, “Fundamentally, most people don’t cheat. But, some do.”
If you’re concerned your candidates could be cheating, try typing your interview questions into ChatGPT and see what’s generated. For those interviews where candidates have access to the questions before, pull up your own ChatGPT window and compare the answers.
Try to also just set some ground rules. Assure your candidates that the interview process is an opportunity to get to know them—and that ChatGPT doesn’t provide personalized, authentic content.
Keller suggests that including a simple note could help.
“Remind your candidates you want to see what their brain does, so simply ask them not to use ChatGPT or external resources.”
Be honest with your candidates. ChatGPT can be a great resource for interview prep, but tell them you want to learn about them and what they enjoy, so sharing stories outside of direct work responsibilities like volunteering or organizing a team activity gives an authentic view into who your candidates are.
HireVue is proud to be at the forefront of conversations around ethical AI and ensuring hiring teams are using technology in safe and ethical manners.
However, so much of the discussion about AI can be based on fear and/or misinformation—will AI take over and eliminate jobs? According to Keller, the conversation has changed quickly.
“Even 3 years ago, the conversation around AI was more constrained, with some people scared it would eliminate programmers. That’s not going to happen anytime soon. The best AI usage is a marriage between human beings and technology. For technical employees, a lot of code will be automated in the future. The way programmers work will change dramatically, but engineers aren’t going away. They’ll just need to become more skilled at using new tools. Job responsibilities and the way we leverage tools may change, but AI has been around for decades. How we react to it and use it will become more sophisticated.”
So hiring teams, you can take a deep breath, because AI is here to make your jobs easier. And with the right tools in place, your hiring will be faster, fairer, and more flexible—making it a better experience for you and your candidates.
Have questions about ethical AI or how to make your hiring smarter than ChatGPT? Request a demo.