AI is taking over hiring. What happens when humans disappear? 

Artificial intelligence is being integrated into nearly every aspect of the workplace, including areas where many believe human involvement should remain central. As the technology advances, its use in scouting and hiring is becoming increasingly widespread, reshaping how companies identify, assess and select candidates.

That integration can now span almost every stage of the hiring process, from writing job postings and filtering applications to conducting AI-led interviews, where applicants’ responses are recorded, transcribed and summarized for recruiters to review.

According to research, up to 99% of hiring managers integrate AI into some part of their workflow, with more than 51% saying it significantly impacts early-career and general recruitment. LinkedIn’s 2025 Future of Recruiting report found that 37% of recruiting organizations were experimenting with or actively integrating generative AI.

AI is also being used on the other side of the hiring equation. According to Gartner, 39% of job seekers have used AI somewhere in their application, most often to generate CV text, used by 54% of those candidates, or a cover letter, used by 50%.

THE AI PROCESS

Dr. Mohamed Gomaa Abdallah, Co-Founder & CEO of Daturial, says AI integration in hiring has gone further than most people outside recruiting realize.

“The distinction I keep coming back to is between automating a process and owning a decision. AI now does much of the first. It rarely does the second, and in my view it shouldn’t.”

He says the process goes something like this: An applicant could apply for a role, within a minute, a parser has turned their CV into structured fields and a model has scored it against the job description. A chatbot asks them two eligibility questions, one about work authorization and one about notice period. The next day they get a link to a recorded interview in which an AI asks the questions and produces a transcript and a summary. On day three a recruiter reads the summary and decides whether they get a call. Their first conversation with a person happens on day five.

Gomaa notes that AI tools can now write job descriptions, source candidates from large databases, parse thousands of CVs, rank applicants against a job’s requirements, answer candidates’ questions through chatbots and book interviews. Some run assessments. Some transcribe interview answers and score them.

Gomaa says AI has not yet taken over the decision-making aspect yet.

“We are well past the experimental stage, but we are not at the point where AI runs recruitment from end to end in most organizations. It has taken over most of the top and middle of the funnel, and the accountability around it is still catching up.”

Similarly, Ahmed Saad, Artificial Intelligence Engineer, GameIn, distinguishes between AI being used to automate individual tasks and taking over the entire decision-making process.

“The important change is that AI is moving further down the recruitment funnel. A few years ago, most automation happened at the top of the funnel, where companies might receive thousands of applications and needed help filtering them.”

“Now AI is appearing in stages that were traditionally much more human, particularly interviews, candidate evaluation and recommendations. Technically, it is becoming possible to automate almost the entire process. The bigger question is whether companies should do so.”

Saad cites real examples like Mercor, which uses its AI interviewer Monty as part of a platform that can analyze a candidate’s background, conduct a live conversational interview, and use that information to match candidates with opportunities. Mercor says Monty now conducts around 10,000 interviews every day.

“This has also created mixed reactions from candidates. Some people become frustrated after repeatedly being rejected following an AI interview, and there are public examples of candidates complaining about the questions, technical problems, or feeling that the system did not properly understand their experience.”

DRIVERS OF INTEGRATION

Dr. Latifa Almuqren, AI Consultant and Associate Professor at Princess Nourah Bint Abdulrahman University, believes the motivation behind AI integration is clear: speed, scale, consistency and cost.

“A human recruiter cannot realistically review thousands of applications with the same level of attention. AI can process large candidate pools quickly, identify relevant skills, handle repetitive administrative work, and allow recruiters to spend less time on tasks such as CV sorting and scheduling.”

She adds that AI is particularly strong at identifying patterns and processing information at scale, while humans are much stronger at understanding context.

“A career gap, an unconventional career path, transferable skills, motivation, or potential may make perfect sense to a human interviewer but appear less attractive to a system optimized around historical hiring patterns.”

Similarly, Saad points to scale as the strongest driver. AI can process high volumes of applications quickly, identify candidates who meet objective requirements, organize information and reduce a large applicant pool to something a human team can realistically review. It can also handle repetitive administrative work such as scheduling, answering basic candidate questions, writing summaries and updating recruitment systems.

“There is also a consistency advantage when AI is used correctly. Human recruiters can become tired, overlook information, or unintentionally evaluate similar candidates differently. A well-designed system can apply the same predefined criteria across a large number of applications.”

“In my view, AI adds the most value when it acts as an extremely capable assistant rather than an unquestioned decision-maker. The final interpretation of those patterns is where human judgment becomes important.”

THE RISKS

For all the efficiency AI can bring to recruitment, the technology also introduces questions about what can be lost when human judgment disappears.

Gomaa says the first thing a company loses when using AI is context.

“A CV is a compressed version of a career, and an algorithm sees only what made it into the data. It cannot know why someone took two years out to look after a parent, why a nurse retrained as a data analyst, or that a self-taught developer with no degree writes better code than most graduates.”

“The model also struggles with the person whose potential looks nothing like anyone the company has hired before, because a system trained on past hires learns what past success looked like, and that is a narrow sample.”

Gomaa says the second notable loss is the candidates themselves. He references Greenhouse’s 2026 survey, which found that 38% of job seekers had walked away from a hiring process because it included an AI interview, while another 12% said they would. Gartner’s 2025 survey found that 25% of candidates trust an employer less when it uses AI to evaluate them.

“A company that loses a strong candidate to a poor automated experience rarely finds out, because the candidate simply stops replying.”

And while Gomaa notes that people are biased too, he says automation changes the scale of the impact.

“One biased recruiter affects the people who happen to land on their desk. One biased model affects everyone who applies through it.”

Researchers at the University of Washington tested the issue directly in 2024, finding that LLMs preferred white-associated names 85% of the time and Black-associated names 9% of the time. They preferred male-associated names 52% of the time and female-associated names 11% of the time.

Gomaa also raises questions around privacy, transparency and accountability in AI-driven recruitment.

“What information is the system analyzing, and how long is it kept? Can a candidate challenge a rejection? Does the employer know why the model ranked one person above another? And when an automated decision discriminates, who is accountable: the employer, the software vendor or the company that built the underlying model?”

“For me the ethical threshold is simple: the greater the effect of an AI decision on someone’s life, the stronger the requirements for transparency, validation and human oversight should be.”

Similarly, Almuqren says companies lose “exactly what they are ultimately trying to hire: the human behind the data.”

She notes that a CV is an incomplete representation of a person. Even an AI-generated candidate profile is still a representation. Human interaction can reveal curiosity, adaptability, motivation, communication, and potential that are difficult to reduce to structured data.

AI systems trained on historical recruitment data can also reproduce patterns embedded in previous hiring decisions. Automation can allow such patterns to operate at a much larger scale unless organizations actively audit their systems for fairness and bias.

“Recruitment is also often a person’s first meaningful interaction with an organization. If that entire experience consists of algorithms, automated messages, and AI interviews, companies may become extremely efficient at processing applicants while becoming less effective at building trust with potential employees.”

A WORKING SYSTEM

Saad says the best way to draw the line is to ask whether a task is primarily administrative or whether it materially affects a person’s opportunity. Automation can make the process faster for both employers and candidates without necessarily removing meaningful human judgment.

Saad says AI can also play a significant role in the early evaluation stages, by filtering irrelevant applications, flagging candidates who meet predefined requirements, and organizing information for recruiters. It can also highlight relevant experience, compare candidates against job requirements and summarize strengths and gaps.

“At that stage, however, I would not allow AI to become the final authority. Choosing who gets hired is a decision that can significantly affect both the candidate’s career and the company itself, so it should still involve human judgment.”

Human oversight remains particularly important during interviews and final selection, where recruiters can consider context and challenge AI recommendations.

“That does not mean humans are automatically less biased than machines; they are not. The advantage of keeping humans involved is that they can consider context, challenge an automated recommendation and take responsibility for the final decision.”

For Saad, interviews can reveal qualities that automated assessments may miss.

“They are not simply about technical skills; they are about the person behind them: their character, personality, mindset, and potential. In every interview I have conducted or experienced as a candidate, the most meaningful moments were never only about what someone could do, but about who they were and who they could become.”

“Because sometimes, one genuine character, one determined spirit, and one person with the right potential can move everything forward with them.”

Gomaa similarly says the appropriate level of automation should depend on the consequences of the decision.

“If a wrong outcome is minor and easy to reverse, automate. If it changes someone’s career or livelihood, a person answers for it.”

Gomaa says regulation is also beginning to establish clearer expectations around AI in recruitment. The EU AI Act classifies systems used to analyze job applications and evaluate candidates as high-risk, reflecting the potential impact of employment decisions on people’s careers and livelihoods.

For Gomaa, the future of recruitment is not about choosing between human recruiters and AI, but determining where each can add value.

“The more useful model is AI for scale and humans for judgment.”

He argues that automation’s impact ultimately depends on what recruiters do with the time it saves. If they use that time to process even more applications, the technology simply makes recruitment faster.

“So the test I would apply at any company using these tools is a simple one. Ask what the recruiters do with the time they got back.”