Interview Insights


Webinar insights part 2: Harnessing the power of AI and the future of tech hiring

The Karat Team image

The Karat Team

blog feature image

As technology’s hiring landscape rapidly evolves, the pursuit of top talent is more competitive than ever before. To confidently hire the best tech professionals possible, talent acquisition (TA) leaders are using artificial intelligence (AI) and automation to transform how they source, evaluate, and hire. 

Earlier this week, Scott Bonneau, Karat’s EVP of Product and Operations, led an insightful discussion during our second webinar in the three-part series, “ Harnessing AI and the Future of Tech Hiring.” The event featured Jocelyn Lai, Global Head of Talent Acquisition at Duolingo; Chad MacRae, VP of Talent Acquisition at Tinder; and Elena Stefanopol, Director of Recruiting at Labelbox. Each talent acquisition (TA) leader shared their insights into successfully integrating AI and automation into their recruiting practices. Here are key takeaways from the thoughtful conversation:

Capitalizing on AI and automation in recruiting

The webinar began with a discussion about how Duolingo, Labelbox, and Tinder generally leverage AI and automation tools in the recruiting process.

“When it comes to information management and creation, that’s AI,” said Jocelyn Lai. “When it comes to building experiences, that’s where human interaction comes into play.”

Chad MacRae and Elena Stefanopol also agreed that AI streamline administrative work, making it easier for recruiters to be impactful in their roles by offering a human connection at every touchpoint. 

Reviewing the risks and rewards

The panel also discussed the risks and rewards associated with integrating AI and automating with hiring. MacRae stressed the efficiency and scalability that AI brings to their recruitment process; however, he also acknowledged the need to be cautious about algorithmic bias and emphasized the importance of human oversight and refining AI models.

“There is no way to do what we’re doing without our sidekick, AI,” he exclaimed. “But, large language models (LLMs) are only as good as the content put into it. We refine the models to ensure decisions are only made by humans.”

Stefanopol highlighted the importance Labelbox puts on allowing candidates to use AI, specifically ChatGPT, during the interview process. She emphasized training interviewers to know how to use AI in interviews, but said before that training is ever developed or deployed, it’s important to start with the intended outcome you want to achieve.

“We had to take a first-principle approach and balance it,” Stefanopol explained. “We had to ask ourselves: What’s important to us? What are we trying to get out of these interviews? What is the best way to do it?”

Stefanopol said their interviewer training now includes guidance on the importance of listening closely to what the candidate is saying, consider their thought process, and ask follow-up questions.

Evolving approaches to recruiting and evaluating talent

When addressing how AI is changing strategies for recruiting and evaluating talent, the panel was split on how it has changed their processes.

“We were built from AI from the beginning, so it hasn’t changed too much for us,” Lai said. “We still value people who publish, who have strong research and coding skills, and who can work well collaboratively.”

Stefanopol and MacRae articulated a deeper impact due to AI. They discussed how Labelbox and Tinder believe some degree of AI competency extends beyond technical roles. Stefanopol described it as a challenge of identifying the “overlap between what you do and what you hire for” — meaning that Labelbox is working to establish a universal technical hiring bar across its entire staff — from sales to recruiting to customer success.

MacRae agreed, and emphasized that regardless of role, all Tinder team members must have a general level of AI proficiency. 

“When we think of AI talent, we think of machine learning and data science, but we also think  of our marketers, our product managers, our product marketing people, our HR team, our communications team — all of them need to be understanding AI,” MacRae stated.

Exploring company policies regarding use of ChatGPT in interviews

The panel also shared differing views on a candidate’s use of generative AI tools, such as ChatGPT, in interviews. MacRae, Stefanopol, and Lai all agreed that generative AI is a great tool for candidates to use to prepare, but disagreed on whether or not it should be used during the interview itself.

“Knowing that people use it, there’s really no other way than to embrace it,” Stefanopol remarked. “We make sure our interviewers are prepared for whatever situation they may find themselves in. I think we’re in a good place to be flexible, so even if they use it in a phone screen, our interviewers are prepared to work with that, so it’s been working for us to just welcome it.”

While MacRae noted that Tinder objects to using ChatGPT for word-for-word responses to interview questions, Lai said Duolingo hasn’t seen widespread use of AI because it doesn’t align with their culture, mission, or talent strategy. Her company also discourages use because of its focus on technical excellence.

The approach of assessing candidates’ technical acumen independent of AI is something that resonated across the majority of engineering and talent leaders in the most recent Tech Hiring Trends Report.

Lai added that because Duolingo is so focused on hiring candidates for skill, they haven’t seen wide use of AI during interviews.

“My hypothesis is we are quite well known for hiring and interviewing people for their technical craft,” she said. “So, people go into interviews to show off how good they are.”

Protecting data privacy and proprietary information

In the rising age of AI, all of our panelists said their companies take significant steps to protect private data and proprietary information. When each organization does use generative AI as a part of its hiring strategy, it partners internally with IT, security, HR, and legal teams to ensure data is protected and best practices are in compliance with laws and regulations around the world. 

“We’re not putting any confidential information within the large language models that can help it train, and we’re also making sure that we’re using ChatGPT4,” MacRae explained.

Stefanopol mentioned that Labelbox’s hiring practices include reviewing requests for the use of ChatGPT and executing company policies that call for a review process that ensures alignment with data privacy requirements worldwide.

Optimizing recruiting with automation tools

Labelbox, Tinder, and Duolingo also leverage AI to automate different steps of recruiting and hiring. Stefanopol, MacRae, and Lai reiterated that these efforts have not come at the cost of human connection.

“We created a whole career framework that included mapping out all of the roles in our company,” Stefanopol exclaimed. “It really helps with recruiting and having scorecards ready so we can connect the dots in the candidate experience.”

Stefanopol also described how this approach has made it possible to automate things like scheduling, an effort that Duolingo has also adopted. Lai went on to say that Duolingo’s scheduling automation makes the most of the talent acquisition team’s time. She also used it as an example of how to keep the candidate experience and quality top of mind.

“I also truly believe that sometimes we forget quality and understanding what that bar looks like is actually even more efficient than just automation,” Lai said. “We really lean in hard on what great looks like from a candidate profile and that saves way more time.”

MacRae shared Tinder’s efforts to build robust interview panels with an eye on improving note-taking processes to create interview scripts.

“We’ve done a lot of work on building out our leadership competencies and our core and functional competencies across every skill set that we hire at Tinder and Match Group,” he said. “We’ve built out robust interview panels and robust questions so panel members and the hiring managers show up on a zoom and be present.” 

MacRae clarified that the next step is exploring tooling that can create interview scripts and then integrate those scripts with their scorecards.

Looking to the future

Looking ahead, Lai envisioned AI tools that could help recruiters with content marketing, ensuring a consistent tone of voice in communications.

“I truly believe talent brand is becoming more important as recruiters become content marketers,” she predicted. “But, it’s really hard for recruiters to write copy in a specific tone of voice, so AI could help us with that.”

Lai also stressed the need to balance automation with the development of essential skills — a need that MacRae shared. He also anticipated a shift in the profile of recruiters, focusing more on emotional intelligence (EQ) as AI tools take on administrative tasks. 

“It’s going to shift the entry-level roles substantially,” he exclaimed. “What you’re really going to have to hire for is EQ to be able to use tooling and understand how to manipulate some of the data sets. I look forward to talent teams with strong EQ that can deliver great candidate experiences and can automate a lot of the administrative burden that we have had in Talent.”

Finally, Stefanopol hoped the future will usher in an era of resources that could help recruiters navigate the ever-expanding landscape of recruiting tools effectively, simplifying the process.

“My wish would be to find something to really help me connect the dots between all of the different tools that exist. This has been the biggest challenge for us to figure out what works for us and what we should be signing on for because everything changes so quickly,” she noted. “We have a lot of options, so it really becomes about what works for us, for my team, my company — what I’m trying to do.”

Bringing it all together

The integration of AI and automation in talent acquisition is not just a trend; it’s a necessity to stay competitive in today’s tech talent landscape. As the discussion revealed, it’s crucial to strike a balance between automation and human touchpoints, protecting data privacy, and adapting to changing recruitment landscapes.

If you want to dive deeper into these insights and learn more about the future of tech talent acquisition, be sure to join the final webinar on AI next month. Sign up now to stay at the forefront of AI-powered hiring.

Ready to get started?

It’s time to start
hiring with confidence

Request Demo