Recruiting and talent acquisition is a top priority for employers, but recruiting well can take immense amounts of human capital. The average employer spends around $4,000 and takes more than three weeks to recruit each new hire.

So it’s not surprising that employers are excited at the possibilities that AI offers for transforming recruitment. The hype around AI suggests that fully automated recruitment tools are the wave of the future. But what’s the reality when it comes to AI’s capabilities and the role it can — and can’t — play in hiring?

Here are a couple of ways that AI is going to transform recruiting, and one way it’s not.

By Using Automation to Find High-Quality Talent

Chatbots are becoming a common means of recruiting talent. These AI-based tools are designed to identify and contact viable candidates, taking much of the legwork out of recruitment. They work across multiple platforms, and often mimic organic human interactions.

These systems make recruiting far more efficient. As HR analyst, author and podcaster Ben Eubanks says, “The number one challenge of recruiting is finding talent and the high-quality people that we need to fill our jobs.” AI-based chatbots can automate and streamline this process, saving employers significant time and money.

By Improving Job Performance for Recruiters

In addition to talent sourcing, AI systems can learn from people. They can be programmed to analyze top recruiters to determine what practices are most effective in identifying and recruiting talent. AI systems constantly learn and improve, which means recruiters can move on to other things as more tasks shift to the AI.

By Creating a More Inclusive Hiring Process

AI-based systems can generate a larger pool of candidates than traditional recruiting does. A properly designed and maintained AI isn’t biased by previous hiring decisions, so it doesn’t continue to replicate the same type of candidate. It’s simply trained to identify the most qualified candidates.

Unfortunately, it’s not entirely possible to de-bias the process, says Vin Vashishta, chief data scientist at Pocket Recruiter. “You can only de-bias your data set if you are doing very, very obvious machine learning, because then the biases that you'll discover are very, very obvious and you can remove them,” Vashishta says. “If you are doing this next-generation machine learning, the biases that are within the data are going to be things that you can't expect, things that you can't anticipate.”

But there are ways to take advantage of AI recruiting without replicating bias. First of all, it’s important to be aware of potential concerns. Vashishta suggests working within AI’s programming to remove bias. Placing focus on data to include in a talent search--as opposed to exclusion factors--can offset a program's initial biases.

By Augmenting — Not Replacing — Humans

AI offers a substantial boon to recruiters, but it can’t replace the human element. “Humans are naturally terrible at analyzing large amounts of data,” Eubanks says. “Even if you're really great as a recruiter, the data show that we're just not very good at sorting through lots of data.” But while AI makes the recruitment process more efficient, it’s important to remember that AI assistance in recruiting is just that — assistance.

AI’s real value isn’t automating candidate searches but in augmenting the things that make us human. It can streamline the recruitment process, but it can’t replace human creativity. “Machines run an algorithm and then they stop, and they're done, and they don't think back over their work,” Eubanks says. “They're not great at creating connections to people and finding creative solutions to problems, but that is something that a human does really well.”