“We think of ourselves as customer zero for our LinkedIn products,” says Teuila Hanson, chief people officer at LinkedIn. Every new AI feature that’s developed for customers is used by LinkedIn’s internal recruiters, who are encouraged to provide feedback for those products. 

The advancements of generative AI and other systems are changing how humans work today, the roles that they will be hired for tomorrow, and how employers will recruit and assess that talent pipeline in the years to come. Nearly two-thirds of Americans believe AI will have a “major” impact on work in the next two decades. And about a fifth of all workers are in jobs that are deemed “highly exposed” to AI, meaning those roles will either be drastically changed or completely eliminated by the disruptive tech.

New AI tools have emerged to help recruiters assess résumés, refine language for job postings, and train talent, and could potentially reduce bias in hiring—that is, if the systems that are built aren’t introducing the discrimination that exists in humanity.

But HR experts say standards should be set to ensure AI is used responsibly for hiring, cross-functional work that should include technology leaders, as well as functions like legal, compliance, finance, and HR. Efficiency gains are important, but don’t work too quickly to remove humans from the process, people team leaders say, and acknowledge to your workforce that AI is going to change how we work.

One feature LinkedIn recently launched is an AI recruiter tool that uses natural language processing to help recruiters refine their searches for talent. If, as an example, a recruiter is looking for a journalist with experience covering Fortune 500 companies, with a focus on diversity and inclusion, the AI-powered search result can pull together some top talent targets for that type of open role. 

“As the models get smarter, I think we are going to see some really cool opportunities in recruiting,” says Hanson. By making recruiters more productive, they can spend more time talking to candidates and really understanding what gets talent excited about the work they do.

LinkedIn also debuted an AI-powered coaching chatbot that uses Microsoft’s Azure OpenAI to process questions and responses. Questions that the chatbot can help address include: “How can I delegate taste and responsibility effectively?” or “How can I lead better meetings?” 

Intuit finds itself in a similar position to LinkedIn’s. Intuit creates AI features for the company’s tax products like TurboTax and QuickBooks, but also uses AI internally.

“If it’s right for our customers, it’s right for our employees,” says Humera Shahid, chief diversity, equity, and inclusion officer and vice president of talent development at Intuit. “We know AI will change the way that we work for the better.”

Within weeks of the generative AI boom, Intuit set up a pilot studio available to around 500 employees, and later made available companywide, to encourage teams to move faster and become more productive. All employees that wanted to access Intuit’s GenStudio had to participate in mandatory training. And the experimentation environment is firewalled so Intuit employees can test and learn without worry of proprietary information being exposed. 

“We want to use generative AI specifically to help our employees be able to not only work better and do their jobs, but also get access to information more easily in their employee experience,” says Shahid.

And while Intuit says it doesn’t have all the right answers on AI yet, it believes the best message to convey to employees is that AI development is a journey. “There’s a lot more leaning in, I think you’ll find, at Intuit than you may find at other companies just because of how our products have evolved,” says Shahid.

Online employment marketplace ZipRecruiter says AI can be especially helpful for automation of repetitive tasks such as writing a job description, employee communications, onboarding planning, and on-site meeting agendas. AI helps sort through hundreds, and at times thousands of résumés, and float the best ones to the top. 

“The faster you can write a job description and get it out, it brings out the opportunity for you to make that human face-to-face connection on an interview faster,” says ZipRecruiter Chief People Officer Amy Garefis.

Garefis says the best way to use AI is to partner the technology with humans. And before any step is made to fully automate a step in the hiring process, ZipRecruiter has to feel comfortable about responsibly replacing work that was previously done by people. And because AI is here to stay, job seekers have to find ways to work within the system.  

Bias, of course, is a concern. Humans have bias, so using an automated tool could potentially introduce new bias in the hiring process. “But I think we also have the advantage of starting from scratch and building systems and making sure that the AI that we build or use does as much as we can—and potentially more—to eliminate bias from a process,” says Garefis.

The government is also starting to emerge as a player in regulating the use of AI in hiring. New York City last summer passed a law to protect workers from bias in hiring. California is mulling a law that would allow applicants to opt out of automated decision-making tools. The Biden administration’s recent executive order directed the development of “principles and best practices to mitigate the harms and maximize the benefits of AI for workers.”

“As AI continues to evolve in the workplace, let’s not forget that it’s more important than ever for business and talent leaders to reinforce that the future of work is still human,” says Hanson.

“We’ve been very careful,” says Marlon Sullivan, chief human resources officer at Johnson Controls, which supplies technology that helps make buildings more sustainable. “We do want to have all of our employees engage with generative AI and be comfortable with it.” 

Johnson Controls’ use of AI is already fairly broad. AI is used to measure carbon emissions for customers and perform predictive analytics to better support technicians in the field. Employees use a version of OpenAI behind the company’s firewall, helping create an environment where AI use is safe, risk-free, and easy to use. 

For several quarters, Johnson Controls has shared an employment survey to get steady feedback on the work environment, engagement with managers, and trust across the organization. AI helps identify themes from those surveys and shares feedback to managers on what they can do differently as it pertains to coaching, better time management, or how to share strategic priorities with colleagues. 

“The AI element of that survey itself has been very instrumental and we have gotten a lot of positive feedback,” says Sullivan. Each survey has an employee net promoter score, which has “increased meaningfully” because AI is providing data that allows managers to make more informed decisions.

Johnson Controls also uses a virtual avatar embedded on the company’s career site to help answer questions from candidates. Over 5,000 candidate interactions have occurred with the AI avatar.

Heading into 2024, Johnson Controls is working to unpack how it can leverage AI, not just through tools that better enable innovation, but that can improve collaboration and bring teams together in novel ways that haven’t been done before. 

“It’s really about how do we leverage AI to better empower our employees, where they are able to reach their full potential and what I call, thrive,” says Sullivan. “It’s not just about being happy that the company knows me, has engaged with me, and that I have a good manager. It’s about, ‘How do I be the best version of me, at a company that I love, and know that the work I do is really noble and impactful?’”