Artificial Intelligence is revolutionizing HR, going to replace recruiters, and is completely neutral... 

That’s a lot of conflicting information about AI! So, to separate fact from fiction in this article, Noota is talking with Indeed, through Anaïs, to explore all the myths and realities surrounding AI in recruitment.

And who better to provide clear insights than Eric Gras, Head of Talent Intelligence?

With over 25 years of experience and a lifelong passion for HR, as he likes to say, Eric has been observing and analyzing employment market trends for the past 7 years using the millions of data points collected daily by Indeed.

His focus: matching candidates with job openings.

His mission: to make AI more useful, more accessible... and above all, more human.

Discover what AI really changes for recruiters here, as well as the most common myths and how they're debunked.

AI in Recruitment: A Revolution or Deja Vu?

It's tempting to think that AI is a brand-new innovation.

And that's understandable. With the arrival of generative AI, a kind of fascination (and even concern) has emerged regarding artificial intelligence.

But if you think this is an unprecedented technological shift, you should know that in the field of recruitment, AI is nothing new. It's been operating for decades, sometimes behind the scenes, integrated into the everyday tools used by recruiters.

Furthermore, Indeed has been using AI in various forms since its inception 20 years ago. It's not a recent change but a core part of their model.

And yes! You might not realize it, but when you browse Indeed, you're directly or indirectly using artificial intelligence, particularly machine learning, deep learning, and LLMs (Large Language Models, an AI model trained on massive amounts of text to understand, generate, and rephrase natural language).

This job board collects the equivalent of 300 terabytes of data every day, which is 20 times the size of the National Library of France, thanks to its large-scale operation (350 million visitors and 30 million job postings, no less!). And the tool doesn't just analyze resumes. It learns from candidate behavior (clicks, scrolling, time spent, geographical preferences, etc.), with 320 data points continuously monitored.

Indeed's stated goal is to offer a solution that serves people, facilitating and improving the reliability of connections, while leveraging technology.

'We collect a huge amount of data, which allows us to understand both the 'behavior' of candidates on one side and the behavior of recruiters on the other. We then determine what works and what doesn't, to try to bridge these two worlds that don't always speak the same language.' Eric

The real novelty of recent years for AI lies more in its increased accessibility through tools like ChatGPT or Noota.

Whereas in the past, only technical profiles were users, AI is now mainstream. It has become tangible, interactive, and can be integrated into all stages of recruitment, from writing job ads to analyzing applications.

In other words, AI isn't new... it's its use that has become more widespread.

Will AI Replace Recruiters... Or Not?

A study conducted by Indeed and Censuswide reveals that over 80% of recruiters surveyed already use AI in their practices. Even better, over 70% of HR professionals believe it improves their productivity, and 69% say it allows them to focus on what matters most – human interactions.

Yet, despite this widespread adoption, a persistent belief remains: that AI will eventually replace recruiters.

In the same study, when HR and recruitment professionals were asked what worries them most, 49% answered: the loss of human judgment in recruitment decisions.

Worse, over 30% fear that AI will replace certain jobs.

'AI doesn't replace recruiters, but it does replace basic and repetitive tasks (screening, resume sorting, scheduling, interview invitations, interview summaries) that account for up to 70% of a recruiter's time!' Eric

The challenge then is to become an 'augmented' recruiter by freeing up, thanks to AI, more time for interviews, qualitative analysis, and relationship building. In the long run, recruiters will be replaced not by AI, but by those who know how to use it, because AI cannot match emotional intelligence and human expertise, according to Eric.

Ultimately, won't AI transform jobs more than it destroys them?

It changes the nature of the tasks performed and pushes for skill development, without eliminating the jobs themselves.

Eric draws a comparison with call center agents. Chatbots have been implemented to handle basic inquiries, so agents can focus on more complex and higher-value tasks, becoming 'augmented agents.'

Small Businesses, Big Power Thanks to AI

Among the misconceptions circulating is that only large organizations can afford to invest in artificial intelligence. However, there are many AI tools available for free or at low cost, including for very small businesses.

Free versions of ChatGPT, Applicant Tracking Systems with integrated AI, ad or message writing assistants, matching tools, or solutions that transform your oral conversations into structured summaries like Noota: there's a wide range of options adaptable to all budgets.

For Eric, while AI doesn't necessarily require large financial resources, it does require an attitude: that of 'test and learn.'

'Small and very small businesses are often the most agile. They test, adjust, and move faster. The important thing is to experiment and measure the real impact on performance.' Eric

Automate, Recruit, Discriminate?

Is AI neutral?

Behind its apparent objectivity, it can actually reproduce and even amplify human biases. Eric emphasizes that artificial intelligence is not magic. It needs to be trained, guided, and supervised.

It learns from existing behaviors and data. And if those are biased, its decisions will be too. A poorly configured AI can then unintentionally discriminate.

The OECD, which plays a role in economic analysis and recommendations and with which Indeed regularly collaborates, confirms this through various studies. Women are the primary victims of algorithmic biases if they are not corrected.

'The risk is freezing historical practices, while a well-used AI can actually open up possibilities in recruitment. AI is a decision-making aid, not an infallible prediction tool.' Eric

Eric illustrates this point with a telling example: personality tests. Their results depend heavily on the candidate's state of mind at the time of the assessment.

Let's take two profiles with equal skills. One is employed, confident, and well-supported; the other has just experienced a layoff, has been looking for a job for several months, in an uncertain environment. Their attitude towards the test will not be the same. However, both potentially have similar chances of succeeding in the targeted position.

This is where the human makes all the difference: an attentive interviewer will be able to contextualize the evaluation, while the machine remains blind to the nuances.

When Human Intelligence Complements Artificial Intelligence

Myths die hard in recruitment, and AI is no exception.

Key takeaway: good recruitment is not just based on a matching score, but on listening, fine understanding of contexts, and personalities.

Eric emphasizes this subtlety with a concrete example from the temp agency world:

'What distinguishes a good recruiter from another in an agency is their knowledge of the local landscape. They know which client to send which profile, not just based on skills, but because they anticipate the human dynamic: this candidate, who is a bit sensitive, will need a supportive manager. No algorithm can yet achieve this level of relational finesse.'

AI requires you to stay in control. Like a GPS, it guides you based on what it knows, and it's the recruiter who's behind the wheel.

AI: What About Critical Thinking?

Matching AI works by making suggestions, based on browsing history, behaviors, and preferences. A bit like Netflix, which recommends movies in a loop based on your initial choices...

This is exactly what you need to avoid in recruitment. Because an AI that relies solely on your past habits can unconsciously narrow your horizons: always the same profiles, the same backgrounds, the same answers.

The bonus tip, according to Eric, is to frequently challenge the tool:

• By testing multiple solutions for the same type of position,
• By comparing the results in terms of quality and diversity,
• By maintaining an ROI-driven approach: measuring what works, what really saves time, or what broadens the spectrum of talent encountered.

The pitfall, to conclude, as you've understood, is to lock yourself into a biased recruitment pattern, like getting stuck in a single category of movies on Netflix.

Want to go further? Find all our resources dedicated to AI for recruiters on our blog.