I want to start with something that sounds contradictory, and then spend the rest of this post explaining why it isn't: AI is one of the best tools available to job seekers right now, and it is also, in its most common use, actively making your job search harder.
Both of those things are true at once, and the difference between the two comes down entirely to how you use it.
The bit nobody's saying plainly enough
Here's the situation as it actually stands. Application volume per job posting has roughly doubled since 2022, and somewhere between 40% and 80% of job seekers are now using AI tools to draft their CVs and cover letters, with some going as far as setting up bots to submit thousands of applications automatically. The result is that recruiters are being handed enormous stacks of polished, keyword-optimised, near-identical documents that tell them almost nothing genuine about the person behind them.
This matters for a reason that's easy to miss when you're the one sending the application rather than the one receiving it: the person on the other end is very often still a human being, and that human being is drowning. Recruiters aren't sifting your CV with infinite patience and unlimited time. They're doing it in the gaps between everything else on their plate, at speed, under pressure, whilst wading through a pile that's grown far faster than their capacity to read it properly. Every application that sounds exactly like the last fifty adds to that pile without helping them find anyone. It slows the whole system down, for everyone, including you.
So when I say "don't use AI to write your job applications", I'm not being precious about it, or clinging to some old-fashioned idea that typing your own words is morally superior. I'm making a practical argument: doing it that way has stopped working, precisely because so many other people are doing it too.
Why letting AI write the whole thing tends to backfire
It makes you sound like everyone else. This is the core problem, and it's a structural one rather than a matter of prompt quality. When thousands of candidates ask a language model broadly similar questions about a broadly similar job, they get broadly similar answers back. Career advisers at Buffalo's College Recruiter have described this openly as an era of "AI spam", where recruiters are flooded with applications that all sound the same because they were built from the same basic prompts. If your cover letter reads like a corporate manual and your CV is a mirror image of the job description with no personal texture to it, you're not standing out. You're disappearing into the noise.
Recruiters genuinely can tell, more often than you'd think. Research reported by the Harvard Business Review found that 68% of hiring managers say they can identify AI-generated content, and separate 2026 data suggests that around 49% of hiring managers will reject a resume they believe was AI-generated outright. Large employers have also started deploying dedicated AI-detection software: SHRM's 2026 hiring technology survey found that 43% of large employers now use these tools as part of their screening process. Detection isn't perfect, and it works far better on fully AI-generated text than on something you've genuinely edited and made your own, but the idea that a polished chatbot draft will simply sail through unnoticed is, for a growing number of employers, no longer a safe assumption.
It hands your own story to a machine that doesn't actually know it. A language model doesn't know what it felt like to solve the specific problem you solved, what you learned from the project that went badly, or which achievement you're quietly proudest of. When you skip straight to "write me a cover letter", you're asking a system with no access to any of that to invent a version of you, and inventing tends to default to vague, safe, unverifiable language. One recruiter-facing piece put this rather bluntly, pointing out that a claim like a vague, unverifiable percentage improvement is exactly the kind of phrasing that makes recruiters sceptical of everything else on the page, not just that one line.
It quietly costs you at interview stage, too. If you've never actually articulated your own achievements in your own words, the interview is where that catches up with you. You'll be asked to expand on something your cover letter claimed, and if you can't do that fluently because you didn't really write it, the gap between the document and the person shows.
It adds to the exact problem you're trying to solve. This is the part that I think gets lost. Every fully AI-written, unedited application adds to the volume overload that's making recruiters slower, more cautious, and more inclined to skim rather than read properly. You're not opting out of a broken system by using AI to do the whole job for you. You're one more contributor to why it's broken.
None of this means AI is the enemy. It means the version of AI use that's caused the problem, "generate my entire application", is not the version worth using.
The honest positives and negatives
It's worth laying this out plainly, because the picture isn't one-sided in either direction.
The case for using AI (done well):
- It removes blank-page paralysis, giving you a starting structure rather than a terrifying empty document.
- It's genuinely good at spotting keyword gaps between your CV and a job description, which matters more than most people realise given how much of the process still runs through automated sorting and screening.
- It can help you organise a job search properly: tracking deadlines, follow-ups, and which version of your CV went where, which matters enormously once you're juggling more than a handful of live applications.
- It's a useful sense-check for tone, clarity, and whether a paragraph you've written actually reads the way you intended.
The case against using AI (done badly):
- It produces generic, interchangeable output when used to write from scratch, and that sameness is precisely what gets filtered out in a flooded market.
- It risks fabricating or exaggerating achievements you then can't back up under questioning.
- Detection tools and sceptical human reviewers are both a real and growing risk, not a hypothetical one.
- Over-reliance means you stop developing your own ability to describe your own work clearly, which is a skill you need for the entire rest of your career, not just this application.
The case for not using AI at all:
- Your voice, your specific details, and your real judgement come through naturally, which is exactly what a tired recruiter is hoping to find in the pile.
- You build and sharpen the skill of self-advocacy, which pays off well beyond this one job search.
The case against avoiding AI entirely:
- You lose a genuinely useful thinking partner for structure, keyword alignment, and organisation.
- You spend far more time on formatting and administrative tasks that a tool could handle for you in seconds, time you could instead spend on the parts that actually need your judgement, like deciding which specific stories to tell.
- Blank-page paralysis is real, and starting from nothing is often where people give up or produce weaker work than if they'd had a scaffold to react to.
Put simply: the problem was never the tool. It's the instruction you gave it.
When AI genuinely helps, rather than just helping you finish faster
The best use of AI in a job search isn't as a ghost-writer. It's as a thinking partner, an editor, and an organiser, three roles it's actually well suited to.
Use it to think, before you use it to write. Ask it questions about the job description, what a phrase in the listing is really asking for, which of your experiences might match, what a recruiter in that sector typically cares about. This is where AI adds real value: helping you understand what's being asked, not answering on your behalf.
Use it to structure, not to author. Ask for an outline, a running order, a list of the sections a strong application in this field usually contains. Then write the substance yourself, into that structure.
Use it to check, not to create. Once you've written a draft in your own words, that's the moment to bring AI back in, to check tone, catch generic phrasing, flag anything that sounds vague or unverifiable, and tighten language that's gone on too long.
Use it to translate and reframe, not to invent. If you did something technical and need to explain it to a non-technical hiring manager, or did something in one industry and need to explain its relevance to another, that reframing is a genuinely strong use of the tool, because the underlying fact is still yours. AI is just helping you say it more clearly to a different audience.
Use it to organise the process around the application, not the application itself. Tracking what you've sent where, when to follow up, and which version of your CV matched which role is administrative work AI handles well, freeing up your energy for the bits that actually require you.
The 10 prompts, if you're going to use AI
Every one of these is written to keep you as the author. None of them ask AI to produce a finished cover letter or CV on your behalf. Paste in the job description and your existing CV or notes where indicated, and adjust the bracketed details for your situation.
1. Requirement-matching, not writing
"Here is a job description: [paste job description]. Here is my CV: [paste CV]. Do not write anything for me. Instead, list the five requirements in the job description that most closely match my background, and for each one, ask me a specific question that would help me recall a real example I could use to demonstrate it."
2. Keyword gap check
"Compare this job description [paste job description] with this CV [paste CV]. List any important skills, tools, or terms that appear in the job description but are missing or underemphasised in my CV, so I can decide which ones genuinely apply to me and add them myself."
3. Structure, not substance
"I'm writing a cover letter for [job title] at [company name]. Give me a suggested paragraph-by-paragraph structure and what each paragraph should achieve, but don't write any of the actual content. I'll write the content myself once I have the structure."
4. Turning a rough note into a clear point
"Here is a rough, unpolished note about something I achieved at work: [paste your own rough notes]. Help me tighten this into one or two clear, specific sentences, keeping every fact exactly as I've described it and not adding any detail I haven't given you."
5. The "does this sound like me" audit
"Read this paragraph I've written for a job application: [paste your own draft paragraph]. Tell me honestly whether it sounds generic, overly formal, or like it could have been written by anyone, and point out specific phrases that feel vague or like AI-style filler, without rewriting it for me."
6. Translating technical work for a non-technical reader
"I did the following in my job, described in my own words: [paste your own description]. I need to explain this to a hiring manager who isn't from a technical background. Help me rephrase it in plainer language whilst keeping every fact accurate and nothing invented."
7. Preparing for likely interview questions
"Based on this job description [paste job description], what are the ten questions I'm most likely to be asked at interview, and for each one, what is the underlying thing the interviewer is actually trying to find out?"
8. Checking for unverifiable or exaggerated claims
"Read this CV bullet point I've written: [paste your own bullet point]. Tell me honestly whether it sounds vague, exaggerated, or hard to verify, and what kind of specific detail would make it more credible, without writing the detail for me since I need to supply that myself."
9. Tailoring an existing draft to a new role
"Here is a cover letter I wrote for a previous application: [paste your own previous cover letter]. Here is a new job description: [paste new job description]. Point out which parts of my existing letter are still relevant, which parts are no longer relevant, and what's missing, so I can decide what to rewrite myself."
10. Organising your search
"Help me design a simple table I can use to track my job applications, including columns for company, role, date applied, CV version used, follow-up date, and status. Don't fill in any details, just give me the structure."
Notice what none of these do: none of them say "write my cover letter" or "write my CV". Every one of them keeps a decision, a fact, or a piece of judgement sitting with you, and asks AI to support that decision rather than replace it.
Where this leaves you
The job market right now is genuinely difficult, and a meaningful part of why it's difficult is the sheer volume of applications flooding in, much of it generated the same way, by the same tools, in response to the same prompts. Adding to that pile with a fully automated application doesn't give you an edge. It makes you one more indistinguishable entry in a stack that a tired human being still has to work through.
Used properly, though, AI is a genuinely good research assistant, editor, and organiser. It just isn't, and shouldn't be, the author of your own story. That part still has to be you.