The growing role of AI in talent acquisition: A change for the better, a change forever

The AI switch: How experts went from caution to conviction

There’s a lot of conjecture that recruiters avoided using AI solely because they were worried about being replaced. But that’s not the full picture.

Let’s look at what actually made talent acquisition professionals wary in the past and what changes they made to go from cautious to convinced.

Flawed hiring decisions

As Charly Huang, HR expert and senior business advisor, rightly says, “AI is no more intelligent than the criteria we choose.” Early AI algorithms for hiring heavily relied on keyword matching, scanning candidates’ resumes, application forms, and public profiles for specific words and phrases related to the job. 

The problem? Candidates could get wrongly selected – or rejected – based on something as trivial as phrasing. For instance, AI could reject an applicant for using the words “client outreach” instead of “generates leads,” costing you perfectly skilled talent. 

Early AI tools also struggled to understand context. For example, they could shortlist all candidates with the words “manager” or “team lead” on their resumes for a job that’s more about project management than people management. 

What’s helping?

Employers are switching to newer, more advanced AI tools that prioritize job-specific skills over keywords when scanning resumes, applications, interview content, etc. They also keep humans deeply involved in both criteria-setting and decision-making. Huang, for instance, shares: “I personally review results to ensure we do not overlook better candidates or introduce bias.” 

Biases

AI tools are trained on historical data, meaning they can inherit the same human biases that plagued past hiring decisions. I read an excellent example of this in the Harvard Business Review, which exposes how AI tools in job boards perpetuate bias:

“Personalized job boards like ZipRecruiter aim to automatically learn recruiters’ preferences and use those predictions to solicit similar applicants [...] 

“If the system notices that recruiters happen to interact more frequently with white men, it may well find proxies for those characteristics (like being named Jared or playing high school lacrosse) and replicate that pattern. This sort of adverse impact can happen without explicit instruction, and worse, without anyone realizing.”

The scary part is that these biases have also led to AI-related lawsuits, such as Mobley v. Workday, in which Workday’s AI screening tools were accused of discriminating against older candidates. 

What’s helping?

Talent acquisition teams are combating bias by implementing techniques such as “blinding.” Research shows that removing candidates’ names, genders, locations, and other identifiers can minimize bias. This way, neither humans nor AI can be swayed when making hiring decisions. 

What’s more, recruiters have also become more careful about the AI tools they integrate into their hiring process, ensuring that these systems are regularly audited and checked for bias. 

Opacity (the black box issue)

Khurana notes that one of the biggest blockers to adoption has been the so-called “black box” problem, where recruiters can’t always understand how AI makes decisions or whether those decisions are based on relevant data, such as skills and capabilities, rather than bias. 

When this happens, not only do you not know if AI is making the right calls, but you also can’t explain how candidate selections were made in a court of law (if you did get sued). 

What’s helping?

Employers are increasingly choosing AI tools that provide rationale for each recommendation and keep humans firmly in the loop. This gives recruiters clarity instead of leaving them in the dark. 

Khurana puts it well: “We’ve reframed the concern by making explainability a priority and encouraging recruiters to engage in the process.” 

Where we are today: How talent acquisition experts are really using AI

Korn Ferry’s latest research shows that 70% of talent acquisition experts believe AI is making hiring more efficient. Yet, TestGorilla’s own research found a 17% drop in employers integrating AI into their hiring processes in 2024 compared to the previous year. 

What’s happening? 

In my opinion, it’s not that recruiters aren’t using AI; they’ve simply stopped using AI for AI’s sake – i.e., to look more “innovative.” Instead, they’re using it more thoughtfully than before, and it now touches nearly every part of the hiring process. 

And we have the receipts to back this up. We asked employers how they’re using AI in hiring in 2025, and here’s what we found: 

• 51% use it to source candidates

• 60% use it to write job descriptions 

• 59% use it to screen resumes

• 20% use it to interview candidates 

Peter Murphy, CMO and equity partner at Ella Weddings, tells us he also uses AI for candidate engagement.

With that in mind, let’s dive into how AI has made its way into each of these areas. 

Sourcing

Historically, the time and effort required to hire for active roles were immense, so sourcing often took a back seat. Now, with AI in the mix, recruiters can spend more time scouting passive candidates (those who aren’t actively job hunting) and build their talent pipelines for future roles. 

And you don’t need to spend time searching campuses, events, or online platforms like LinkedIn, either. Today, you can opt for pre-vetted sourcing pools and AI-powered recommendations. 

Take TestGorilla Sourcing, for example. It gives you access to more than two million skills-tested candidate profiles and uses AI to match them to roles based on hard data such as skills, location, and more. This gives you fast, qualified matches while reducing bias and false positives and negatives. 

Job descriptions 

A recent report by Boston Consulting Group notes that “content creation, such as writing job descriptions,” is one of the most popular use cases of AI in HR. It also found that 70% of companies already using AI or generative AI (genAI) in HR are doing so for this purpose.

AI tools like ChatGPT apply machine learning (ML) and natural language processing (NLP) to craft job descriptions from scratch. As recruiting strategist Mike Wolford explains in his book The AI Recruiter: Revolutionizing Hiring with Advanced GPT-Powered Prompts, you can enter a job title, skills needed, responsibilities, and even a specific tone, and the system “will respond to such a prompt quickly and efficiently.”

It doesn’t end there, though. AI can also be used to ask for feedback on previous job descriptions, spot and fix biased or discriminatory language, and more. 

Screening

When AI first entered the hiring space, recruiters quickly recognized its value in scanning large volumes of applications and filtering candidate profiles using keyword matching. These tools are now integral to many applicant tracking systems (ATS), and employers have felt the benefits firsthand. 

Lacey Kaelani, CEO & co-founder at Metaintro, shares with TestGorilla: “We use AI for job matching, processing user profiles against our database of 600 million+ job postings […] Obviously, no human could review that volume.“ 

Chris Sorensen, CEO of ARMOR Dial and PhoneBurner, similarly notes that AI tools have “cut out screening time by probably 30–50%.”

But despite this efficiency, traditional keyword-matching tools can be flawed, biased, and opaque. That’s why many talent acquisition teams are switching to more advanced AI screening tools. 

One example is TestGorilla’s AI resume scoring. Instead of relying on simple keyword detection, it extracts skills-based job criteria directly from a role description – e.g., “ability to use Microsoft Excel for financial modeling” instead of just the keyword “Excel.” You can also review and refine these criteria to ensure they’re accurate and context-relevant. 

Once candidates upload their resumes, TestGorilla’s AI automatically removes names, pronouns, locations, and other identifiers to reduce bias. It then scores each resume based on how well it aligns with the predefined job criteria and provides a brief explanation for the rating. Recruiters can override the scores, and they still make final decisions – but now they have clear, structured input from AI to guide the process. 

To top it off, TestGorilla conducts regular audits to make sure outcomes are fair across demographic groups.

Candidate communications and engagement

AI and automation have been game-changing for candidate communications and engagement. For instance, you can set up automatic “thank you for applying” messages and use AI to schedule interviews, guide candidates through assessments, and provide them with feedback. Some employers even use AI chatbots to answer applicants’ questions in the application portal. 

According to Insight Global’s survey, 98% of employers saw significant improvements in hiring efficiency by using AI for tasks such as scheduling interviews, sending candidate progress messages, and delivering rejection messages.

And candidates benefit, too. As Huang tells us, “AI doesn’t just cut time to hire for recruiters, but candidates experience the process as smoother [and] more time-respectful.” This leaves candidates with a positive first impression of your company.

Beyond this, AI can be used to engage passive candidates. According to BCG, a common use of AI in talent acquisition is email marketing: Recruiters can keep talent pipelines warm with personalized newsletters and updates, all without manual effort.

Interviewing 

Initially, AI only assisted with scheduling interviews. Today, it plays a far more strategic role. Instead of relying on phone screening, which is time-consuming and prone to bias, top talent teams now use AI video interviews to screen candidates.

TestGorilla’s AI video interview tool, for example, interviews candidates using predefined, structured prompts, transparent scoring systems, and human review. Simply put, you decide what the AI should ask, step aside for all the manual work, and step back in to review scores and feedback. 

There’s proof that this approach works wonders. A study by the University of Chicago’s Booth School of Business found that candidates interviewed by an AI agent were 12% more likely to get a job offer and 17% more likely to stay with the company. It also concluded that employers cut interview costs by almost half for AI interviews compared to human-led interviews. 

And the best part? Contrary to popular belief, 78% of candidates actually preferred being interviewed by AI. They felt it was lower pressure, less biased, and more convenient. 

Offers and onboarding

AI is also helping talent acquisition teams close the loop by assisting with job offers and employee onboarding.  

You can now use generative AI tools like ChatGPT to produce personalized, polished offer letters that reflect your company’s tone and branding. Additionally, you can set up your ATS to automatically send offer letters and follow up with candidates for their e-signatures using built-in AI workflows that trigger these actions in the correct order. 

Once an offer has been accepted, AI and automation assist with tasks such as data entry, document verification, and scheduling training for new hires – all of which were once the recruiter’s job. That makes onboarding remote workers much easier. 

In addition, one study found that when AI guides new hires and sends them reminders, it helps them remember to complete their onboarding tasks. Overall, this helped reduce onboarding time from days to less than 15 minutes.

Treading carefully: Why AI can’t – and shouldn’t – replace humans

Every expert who spoke to TestGorilla says that despite AI’s incredible benefits in streamlining hiring, it’s humans who need to call the shots. 

For example, Stephanie King, CMO at Octopus Deploy, accepts that AI can “create efficiencies around sourcing, screening, and candidate engagement.” But she doesn’t rely on it beyond a point. “AI is a powerful accelerator in our talent acquisition process, but not a decision-maker,” she says. 

And even within hiring processes that do rely on AI, recruiters are still very much involved. Here’s why:

AI can push some talent away

Candidates still don’t trust that AI can make fair hiring decisions. Research by Pew found that 71% of candidates wouldn’t consider applying to a company where AI is making final hiring decisions. 

The problem isn’t just about trust. Candidates also want to connect with real people to understand more about the company and ask questions. In fact, a recent Express Employment Professionals and Harris Poll survey showed that 84% of job seekers wanted to be able to ask their questions to humans rather than AI. 

It works the other way around, too. According to the survey, 87% of candidates feel AI can’t vet their soft skills and cultural contributions. 

They’re not wrong. For instance, HireVue has had to discontinue its AI facial analysis tool that judged candidates’ personality traits during video interviews. In one article, Merve Hickok, SHRM-SCP, an expert in AI ethics, bias, and governance, correctly points out: 

“Facial expressions are not universal – they can change due to culture, context and disabilities – and they can also be gamed [...] So, accuracy in correctly categorizing an expression is problematic to start with, let alone inferring traits from it.”

AI isn’t great at reading between the lines

Even today’s advanced AI tools struggle to interpret context. Murphy Lewis explains, “We always ask ourselves whether someone has the disposition to thrive in our context, which we believe AI will have a difficult time evaluating.”

For example, a human recruiter might see a five-year career break to care for a loved one as a sign of empathy, patience, and resilience – a nuance that AI can’t capture. 

And Khurana agrees: “AI can find skill alignment and highlight skills, but it cannot replace a recruiter's gut of whether the trajectory of [a] candidate's career aligns with the mission of the company or whether the candidate's values align.”

AI is not made for finding falsehoods and nuance

Today, many candidates use AI to write their resumes and fill out application forms, and these tools have been shown to inflate or flat-out lie about the candidates’ accomplishments. Moreover, TestGorilla research found that among those who cheat in their applications, 70% use AI to do so. 

The point is, AI hiring tools are simply screening for keywords or skills. They can’t quite pick up on these exaggerations the way a human recruiter might. For example, a resume could falsely claim that a candidate has led a marketing project when, in fact, they only assisted. While AI screening tools may not detect the inflation, a human recruiter is more likely to spot it when they review application answers or ask detailed interview questions. 

Talent acquisition is at its turning point: AI powers the engine, but humans steer the wheel

We’ve come a long way since AI first entered the world of hiring. Now, it’s no longer just handling logistics; it's playing a supporting role across nearly every stage of talent acquisition. And the keyword here is “supporting.”

As Marc Sylvester, VP at Connext Global, says, “It’s true that AI is helping recruiters and managers move faster, but speed only works if it’s matched with fairness and clarity.” Recruiters have learned the hard way that this only comes when machine precision is paired with human judgment. 

Contributors

• Charly Huang, HR Expert and Senior Business Advisor

• Chris Sorensen, ARMOR Dial and PhoneBurner, CEO

• Lacey Kaelani, Metaintro, CEO & Co-Founder

• Marc Sylvester, Connext Global, Vice President

• Pankaj Khurana, Rocket Company, VP Technology & Consulting

• Peter Murphy Lewis, Ella Weddings, CMO and Equity Partner

• Stephanie King, Octopus Deploy, Chief Marketing Officer