Yes. Of course they can.
The irony is that it is easy to spot precisely because AI writing is too good. Perfect rhythm. Flawless transitions. Every sentence lands exactly where it should. Human writing does not work like that. It has friction, odd choices, a sentence that runs long because the person got excited. AI produces prose that no real person under any amount of enthusiasm would actually write. That smoothness is the tell.
One morning I opened my inbox to review applications for a product marketing role. The first cover letter began with "I am excited to apply." The second one did too. By the fifth, I stopped counting. Somewhere around the twentieth application, the pattern was clear: I could not tell these people apart. Not because they lacked qualifications, but because something had smoothed away everything that made them sound like real people.
The patterns are always the same. Same openers. Same structure. Enthusiasm with no detail behind it. Nothing that tells me why this person, why this role, why now. Most candidates assume I would just reject it and move on. It was not that simple. Once I spotted it, it changed how I read the rest of the application.
The Flood: What My Inbox Looked Like by the End
In my inbox the shift was dramatic. In early 2024, I would flag maybe one application in ten as AI-generated. By mid-2025, the ratio had climbed to six or seven out of ten. A Resume Now survey of 925 HR professionals found that 90% of hiring managers reported an increase in low-effort or spammy applications, largely driven by AI tools. That matched what I was seeing every Monday morning.
The volume was not the problem, I had always reviewed a lot. What changed was that the variety disappeared. Before, you would get a mix: strong ones, weak ones, ones with odd formatting, ones that were clearly written at midnight. Each one felt like a different person. By the end I was opening twenty applications in a row that felt like the same document with a different name at the top.
When the majority of applications use the same AI models, the output clusters around a narrow band of phrasing, structure, and tone. Instead of a diverse pool of candidates, recruiters see a wall of near-identical text. The candidates who write their own applications now stand out more than ever.
And when a real personality did come through? I noticed it immediately. It had become that rare.
The part most candidates miss: you are not competing against the job requirements. You are competing against a pile of applications that all sound exactly like yours.
The Seven Tells That Give AI Applications Away
These are the patterns that showed up consistently. Not every AI application has all seven, but most have at least three or four.
1. The "I am excited to apply" opener
This is the single most common tell. Most AI-generated cover letters begin with some variation of "I am excited to apply for the [exact job title] role at [company name]." It is technically correct. It was also the opening line of roughly 40% of the applications I reviewed. When I saw it, I already knew what the rest of the letter would look like. The tone was set. The template had announced itself.
2. Parroting the job description back
AI tools are trained to mirror the language of the prompt. When a candidate pastes a job description into an AI tool and asks for a cover letter, the output directly paraphrases the requirements back at the recruiter. "You are looking for someone with strong communication skills and experience in cross-functional collaboration" followed by "I bring strong communication skills and extensive experience in cross-functional collaboration." I wrote the job description. I knew what it said. What I needed to hear was something I did not already know.
3. Generic enthusiasm without specifics
AI-generated applications radiate enthusiasm that floats above any actual detail. "I am deeply impressed by your company's commitment to innovation" could apply to any company on the planet. When a candidate researched the company, they mention a specific product launch, a recent decision that caught their attention, a particular aspect of the culture they connected with. The difference in energy is immediate. AI-generated text stays safely vague because the model does not actually know why this candidate wants this role at this company.
Vague enthusiasm is worse than no enthusiasm. When a candidate writes "I admire your company's mission" without naming the mission, it signals that they could not be bothered to look it up. Specificity is the clearest signal of genuine interest.
4. Perfect grammar, zero personality
Human writing has texture. People use sentence fragments sometimes. They start sentences with "And" or "But." They have verbal habits and rhythms that make their writing feel lived-in. AI-generated text is grammatically flawless and completely flat. Every sentence is properly constructed. Every paragraph transitions smoothly. And none of it sounds like a person sat down and thought about what they wanted to say.
5. Buzzword density that no person would use naturally
"Spearheaded cross-functional initiatives to drive strategic alignment and optimize stakeholder engagement." No one talks like this in conversation. No one writes like this when they are being themselves. But AI models, trained on corporate text, produce this kind of language constantly. When every sentence contains three or four corporate buzzwords, the application reads like a parody of a LinkedIn post, and the person behind it disappears entirely.
6. Identical phrasing across candidates
This is the tell that candidates cannot see but that is impossible to miss when you are the one reviewing the pile. When I reviewed fifty applications for the same role and fifteen of them contained the phrase "I bring a unique combination of," those fifteen applications were written by the same tool. The candidates thought their application was unique. From where I sat, it was one of many copies.
7. The resume echo
AI-generated applications often restate resume bullet points in paragraph form without adding context, motivation, or reflection. The result reads like a reformatted resume, not a letter from a person who has a reason for writing. A human candidate would explain why a particular experience matters for this role, what it taught them, what drew them to apply. An AI-generated application just restates that the experience exists.
Read your application out loud before submitting. If it sounds like a press release or a corporate memo, it probably reads that way to the recruiter too. Your application should sound like you explaining to a colleague why you want this specific job.
What Happened After I Spotted One
Most candidates think spotting an AI application means automatic rejection. It did not work that way for me. What it did was shift how I read the rest.
Once I spotted it, the question was no longer "is this a good candidate." It became "did this person engage with this at all." And that is a different evaluation entirely.
Some applications were clearly AI-assisted but still had something real in them. A specific project. A detail about the company. A sentence that sounded like an actual person wrote it. Those were fine. Using a tool to organize your thoughts and clean up your writing is not a problem. That is just good editing.
The ones that concerned me were the ones where there was nothing left after the AI was done. No personal detail, no company reference, nothing that suggested the person read the job posting before pasting it into a prompt. That told me something. If someone will not spend thirty minutes on their application, what do they do when the work is harder than that?
A recurring theme in recruiter conversations during my last years in the field: AI-assisted applications are acceptable, but AI-replacement applications are not. The line between the two is whether the candidate's own thinking is visible in the final product.
Daniel Chait, CEO of Greenhouse, has called this dynamic an "AI doom loop": candidates use AI to generate more applications, recruiters use AI to filter the growing volume, and both sides lose trust in the process. Average time-to-hire has climbed to 44 days in 2026, up from 31 days two years ago, partly because of this cycle. The Greenhouse 2025 Workforce Report found that 91% of recruiters have spotted candidate deception, and 34% spend up to half their week filtering spam and junk applications.
I never rejected AI-assisted applications automatically. But when two candidates were roughly equal and one clearly wrote their application themselves? That was not a hard decision.
How to Use AI Without Getting Flagged
AI is a useful writing tool. The problem is not using it. The problem is using it as a replacement for thinking instead of a supplement to it. In that same Resume Now survey, 62% of employers said they are more likely to reject AI-generated resumes that lack personalization, while 78% said personalized details are what signal genuine interest and fit.
A practical approach that keeps AI helpful without making your application indistinguishable from every other AI-generated submission:
Start with AI, finish with you. Use AI to generate a structural outline or a rough first draft. Then rewrite every section in your own words. Add the specific project where you demonstrated the skill they are asking for. Name the product of theirs that you actually use. Reference the part of the job posting that made you pause and think, "That is exactly what I want to do next."
Delete every sentence that could apply to any candidate. If a sentence in your application would still be true with a different candidate's name attached, remove it or replace it with something only you could have written.
Break the AI pattern. Change the opening line. Remove the buzzwords. Add a sentence fragment if that is how you naturally write. Let your personality come through the polish, because that personality is the one thing no AI can replicate.
I am excited to apply for the Marketing Manager position at Acme Corp. With over five years of experience in digital marketing and a proven track record of driving growth, I am confident I would be a valuable addition to your team. I am particularly drawn to your company's commitment to innovation and customer-centric approach.
Your Q3 blog series on retention metrics changed how I think about content strategy. I have spent the last three years building the content program at Base Co from zero to 40,000 monthly readers, and the approach your team outlined in that series is exactly the direction I want to take next.
The difference is not about writing quality. Both versions are well-written. The difference is that one could have been written by anyone for any company, and the other could only have been written by one specific person for one specific role. That specificity is what made me lean forward.
If you are applying to many roles and wondering whether volume or quality matters more, the answer is not a number. It is a strategy.
What Stands Out When Everything Looks the Same
When most applications in my inbox read identically, the bar for standing out was surprisingly low. You do not need to be a brilliant writer. You do not need a dramatic career story. You need to be specific. That is it.
The applications that made me pause and pay attention shared a few common traits.
They referenced something concrete about the company. Not "your innovative culture," but "the decision to open-source your design system last quarter." This told me the candidate did research that went beyond the About page, and it told me something about what catches their attention.
They connected their experience to the role with a specific example. Not "I have strong project management skills," but "I managed the migration of 200,000 user accounts to a new platform over six weeks with zero downtime." The second version gave me something to evaluate. The first gave me nothing to hold onto.
They sounded like a person. This was the simplest and most powerful differentiator. When I read an application and could hear a real human voice behind the words, it created a connection that no amount of AI polish can replicate. Imperfect writing that sounds authentic beats perfect writing that sounds generated.
Before you submit, try this test: cover the name on your application and read it as if you were the recruiter. Would you be able to tell which candidate wrote it? If it could belong to anyone, it needs more of you in it.
They acknowledged trade-offs honestly. One of the most refreshing things a candidate can do is acknowledge a gap directly. "I have not managed a team this large before, but here is what I did when I scaled my team from three to eight" is more compelling than pretending the gap does not exist. AI never does this. AI presents every candidate as a perfect match. And after reading hundreds of those perfect matches, honesty is what caught my attention.
The candidates who got interviews were not always the most experienced. They were the ones where I finished reading and thought: this person actually wants this job.
The Real Takeaway
AI applications were supposed to help candidates. And maybe they did, for a while. But at this point they have created a baseline of sameness that makes it harder to stand out, not easier. The candidates who get noticed are the ones doing the thing AI still cannot do: showing genuine interest in a specific role, at a specific company, with real examples that only they could have written.
Your application does not need to be perfect. It needs to be yours.