Roughly half of all job seekers are doing it. Here's what's actually happening, and why the ones getting hired aren't the ones you'd assume.

Ask a recruiter in 2026 whether they think candidates are using ChatGPT to write their applications, and the honest answer is no longer a question of if, it's a question of how much. A Financial Times investigation into hiring trends found that roughly half of job applicants now use AI tools like ChatGPT or Google Gemini somewhere in their application process (Entrepreneur), and separate 2026 data puts active usage even higher: 65 to 70 percent of candidates report using generative AI for some part of their job search, including resume drafting, cover letter writing, and company research (ResumeVera, citing Career Group Companies). This has already reshaped hiring in a very literal, practical way: AI-assisted applications have roughly doubled the volume of applications employers now receive per job posting (Entrepreneur).

What's genuinely interesting, and consistently missed in the panicked version of this story, is that using AI isn't actually the problem employers care about. What gets candidates rejected has very little to do with whether AI touched their application, and everything to do with how it was used.

How Candidates Are Actually Using These Tools

The pattern across every major 2026 survey is remarkably consistent, both in what's being done and in roughly how common each task is:

Rewriting and polishing existing bullet points. The most common single task: taking something a candidate already wrote, often rough or unclear, and asking AI to tighten the language, fix grammar, and make it read more professionally. ZipRecruiter data on recently hired workers found resume drafting the single most common AI task, reported by 23 percent of successful candidates (HiredKit).

Drafting a full cover letter from a job description and a resume. The second most common task, used by 21 percent of recent hires according to the same ZipRecruiter data, and reported considerably higher, 72 percent, in a separate Resume Now and Huntr analysis of job seekers specifically asked about cover letter use (HiredKit; detectiondrama.com).

Tailoring one master resume to match a specific job posting's language and keywords. Rather than writing from scratch each time, many candidates now maintain one detailed resume and ask AI to rewrite specific sections to mirror the exact terminology in a given job description, directly addressing the ATS keyword-matching mechanics covered in our earlier piece on CV mistakes that cause instant rejection.

ATS keyword checking and formatting cleanup. Candidates increasingly use AI tools specifically to check whether their resume contains the right density of keywords relevant to a posting, and to catch formatting issues that might cause parsing problems.

Brainstorming and overcoming writer's block, particularly for cover letters, where many candidates report AI's biggest value isn't the final text at all, but simply getting past a blank page and generating a starting structure to then rework (Jofibo).

Does It Actually Work? The Data Is Genuinely Encouraging, With a Catch

This is where the story gets more interesting than the "AI slop is ruining hiring" narrative suggests. A ResumeBuilder.com survey of over 1,000 job seekers found that of those who used ChatGPT to write resumes or cover letters, roughly 70 percent received a higher response rate from employers, and 59 percent were ultimately hired (SHRM). A separate, more recent figure puts the interview rate even higher: 78 percent of job seekers who used ChatGPT for their resume reported getting an interview (Hirelytica).

The catch, and it's a significant one: these numbers reflect candidates who used AI and personalised the output. Resume Now's AI Applicant Report found 69 percent of job seekers using AI for application materials report higher response rates specifically when those materials are personalised, strongly implying the AI itself isn't what's driving better outcomes, the combination of AI speed plus genuine human editing is (detectiondrama.com).

Why Employers Can Usually Tell, and What Actually Gets Rejected

Here the data draws a genuinely sharp, consistent line. In Insight Global's 2025 AI in Hiring survey, 88 percent of hiring managers said they believe they can spot AI use in an application (phrasly.ai), and a Greenhouse 2026 AI in Hiring Report found 91 percent of recruiters report having spotted AI-assisted candidate material (Hirelytica). But almost every survey asking the follow-up question finds the same thing: it isn't AI use itself that triggers rejection. TopResume's survey of 600 US hiring managers found roughly 20 percent would reject a candidate purely for using AI, while 52 percent said using AI for proofreading or drafting support is entirely acceptable, as long as the final product reflects genuine human effort (HiredKit). Resume Now's data found 62 percent of employers specifically reject AI output that lacks personalisation, framing the real rejection trigger clearly (ResumeVera).

The specific patterns that give away unedited AI output, repeated across recruiter-facing guides, are strikingly consistent:

Flawless but generic language that says nothing concrete. AI is remarkably good at sounding impressive while communicating almost no actual information, producing vague claims that float without specific numbers or context attached (Reed).

Overused AI-favourite verbs and buzzwords. Words like "spearheaded," "leveraged," "pivotal," and "intricate" appear disproportionately often in AI-generated text, to the point recruiters now recognise them as a pattern on sight (Hirelytica).

Robotic, overly symmetrical bullet structure. Every bullet point following an identical rhythm and length is a distinctive AI tell that human writing rarely produces naturally.

No specific tools, vendors, or genuinely odd real-world problems mentioned. AI doesn't know which exact software version you used or the strange, specific obstacle you actually hit on a project, so it defaults to generic descriptions where a real candidate would naturally include concrete, sometimes messy detail (Hirelytica).

Hallucinated skills or achievements. When given a thin prompt, AI models can invent plausible-sounding projects or responsibilities the candidate never actually had, a serious problem since it can lead directly to misrepresentation a candidate didn't consciously intend (Reed).

A disconnect between the resume and the cover letter, or between either document and the candidate's actual LinkedIn profile, when AI-generated pieces were produced separately without being cross-checked against each other.

One hiring manager's comment, shared in a LinkedIn recruiter forum and widely cited in recent coverage, captures the actual complaint precisely: reading an unedited AI cover letter "felt like reading a Wikipedia page" (We Are Tweak), technically correct, entirely impersonal, and forgettable within seconds.

The Nigerian Picture

Nigerian career guidance has moved quickly to meet this reality directly rather than pretending it isn't happening. A 2026 Nigerian HR and career resource offering ready-to-use AI prompts explicitly frames the goal as getting AI to understand "Nigerian hiring standards," with example prompts built specifically around local roles, like asking an AI tool to write a CV for a mid-level HR officer with recruitment and payroll experience "in Nigeria," rather than generic international phrasing (Anthrop Management).

At the same time, Nigerian recruitment-focused writing is warning about exactly the same generic-content problem documented globally, with a distinctly local flavour. A 2026 Nigerian career guide specifically calls out "Dear Sir/Ma" generic templates and stock phrases like "I am a hardworking individual" or "I am dedicated and passionate" as instantly recognisable, instantly discarded patterns, noting bluntly that Nigerian recruiters are trained to spot copy-paste applications in seconds in a job market too competitive to tolerate them (JobhardER). This connects directly to a broader theme we've covered before: generic, unproven claims are consistently what gets Nigerian CVs instantly rejected, whether that genericness comes from an outdated template or an unedited AI draft, the underlying failure is identical.

The practical implication for Nigerian job seekers is genuinely encouraging rather than discouraging: AI tools work well here too, but only when fed real, specific, Nigeria-relevant detail, your actual employer names, your actual NYSC placement, your actual achievements with real numbers, rather than a generic prompt left to invent plausible-sounding filler on its own.

A Practical, Research-Backed Way to Actually Do This Well

Write the raw content yourself first, then let AI sharpen it. The winning pattern across nearly every guide reviewed here is consistent: the candidate supplies their real achievements, tools, and specific moments, and AI's job is improving the language around that real content, not inventing the content itself (Hirelytica).

Never let AI write a cover letter end to end from a one-line prompt. "Write me a cover letter for this job" with nothing else produces close to the same generic output every other candidate using that same lazy prompt is also submitting. Feed it your real resume, the actual job description, and specific details about why this particular role interests you.

Edit for your own natural voice afterward, deliberately removing AI-favourite phrasing. A useful habit some 2026 career guides recommend explicitly: after generating a draft, prompt the AI to rewrite it "in my natural tone, with short sentences, one real anecdote, and no buzzwords," then edit further yourself (Inradius).

Always verify every claim the AI generates is actually true. Given AI's tendency to fill gaps with plausible-sounding invention, treat every AI-produced sentence as a draft claim requiring your own fact-check, not a finished, trustworthy statement.

Know your own CV cold before any interview. Whatever AI assistance went into drafting it, you need to be able to explain and defend every single line in person. An interviewer probing a claim you can't actually elaborate on is a far bigger problem than any AI detector.

The Bottom Line

The honest, research-backed conclusion here cuts against both of the loud, simplistic takes on this topic. It isn't true that using AI to write a resume or cover letter is cheating, nearly every major employer survey confirms the large majority of hiring managers find thoughtful AI-assisted drafting entirely acceptable. But it's equally untrue that AI alone can write a winning application, the candidates actually getting interviews and offers are overwhelmingly the ones using AI as a genuine editing and drafting assistant layered on top of real, specific, truthful content, not as a replacement for having anything real to say. The skill worth building isn't avoiding AI, and it isn't blind reliance on it either. It's the judgment to know exactly where the line between the two sits, and the discipline to stay on the right side of it every single time you hit submit.

This article draws on 2026 hiring and job-search research from SHRM, Entrepreneur, Resume Now, TopResume, Resume Genius, Greenhouse, Reed, and ZipRecruiter, alongside Nigeria-specific career guidance current as of late September 2026. AI tool capability and recruiter detection methods are evolving quickly, and the specific figures cited here should be treated as a current snapshot rather than a fixed standard.