Why banning AI in recruitment probably won’t work (and what to do instead)

Why banning AI in recruitment probably won’t work (and what to do instead) 

Every hiring team is having the same conversation. Applications have surged but candidate quality appears to be declining. Responses seem increasingly polished but frustratingly generic. So, it’s not surprising that teams are asking whether they should just ban AI. 

If candidates are using AI to write their applications, how can you trust what you are reading? And what will their performance be like later in the process when you meet them in person? If you can’t tell who wrote the application, why ask for one? 

But will banning AI even work? 

What candidates really think about AI in job applications 

In our latest research surveying people early in their careers, 34% said they do not expect to be able to use AI at any point during a recruitment process.  

However, 47% expect to be able to use AI for creating their CV and cover letter as well as completing application forms. 4% expect to use it throughout the entire process.  

Nearly half of the candidates you are receiving applications from believe that polishing those applications with AI is an entirely reasonable thing to do. They’re not trying to game the system, but doing what they think is the obvious thing to do. 

1 in 4 candidates will use AI anyway 

We asked directly: if an employer said you were not allowed to use AI to support your application, would you still use it? Three quarters said no. But 25% said yes or maybe. Depending on the volume of applications you receive, that could represent thousands of candidates using AI-assisted content in a process where you have explicitly asked them not to. 

In most recruitment processes, you have no reliable way of knowing who those candidates are. AI-generated text detection tools exist, but they regularly flag innocent candidates and miss some that are actually using it. Relying on detection to enforce a ban is a gamble, and an unfair one for the candidates it misidentifies. 

How high application volumes undermine an AI ban 

The 25% who will ignore your ban are not cheaters. They are overwhelmed, eager to get a new role, applying to multiple roles at once to increase their chances, and reaching for tools that make the process manageable. 

Our research confirms the scale of this. Among applicants who use AI, 44% say they are making more applications as a result. And a majority, 55%, describe themselves as applying broadly rather than targeting only roles they are genuinely committed to. AI has changed how applications are written and how many people make. When producing a well-structured, tailored application takes ten minutes instead of an hour, the obvious move is to apply to more roles. 

Banning AI from the application stage does not fix this. It just adds friction, mostly for the candidates who comply, while doing very little to filter out the ones who do not. 

Why AI-polished applications make hiring harder, not just different 

Most hiring teams don’t care that candidates are using technology – most organisaitons want tech-savvy people in their business. They care that the outputs are harder to evaluate. A confident answer to “why do you want this role” tells you less when it has been drafted, revised, and polished by a language model that has read thousands of examples. It’s a lot harder to find out who has genuine motivation, relevant capability, and the potential to succeed. 

Don’t put a blanket ban on AI. Instead design a process that gives you a genuine read on candidates regardless of how they prepared. 

Building recruitment processes that do not rely on an AI ban 

The best processes stop relying on the CV or application form as their primary source of insight. 

Conversational assessments, for example, are significantly harder to prepare AI-generated responses for in advance. Questions are dynamic, follow-ups are unpredictable, and responses need to be produced in real time making a prepared script useless. A candidate who has rehearsed a polished answer to “tell me your greatest achievement” is less prepared for a follow-up that asks them to describe specifically what they found hardest about it, and why. 

It also helps to be honest with candidates about how AI is used and detected in your process, and what you are really trying to assess. When candidates understand that you are looking for genuine thinking rather than a polished answer, most will give you exactly that, especially when the process feels worth engaging with properly. 

Why AI bans send mixed messages to candidates 

We asked whether an employer saying they do not allow AI would make the company more or less appealing. For 57% it would make no difference. But 15% said it would make the employer a little or a lot less appealing.  

For candidates who are already careful about how they use AI, a blanket ban may feel like distrust. For the ones who find applying without AI genuinely difficult, it raises a whole other concern. 

A policy that says “we’ve designed our process to assess you fairly and authentically” sends a very different message than one that says “we don’t allow AI.”   

The bottom line 

Banning AI in recruitment is a policy that most candidates say they will respect, but a meaningful minority will not, and realistically you probably won’t be able to tell the difference. Worse, it does not solve the underlying problem, which is that many traditional screening tools have become far less reliable as a way of assessing real capability and motivation. 

The most effective response is to design screening that’s harder to game. Make it reveal real thinking. It may take a little longer than banning AI but it’s also the only thing that works.