Effort Is the Point: What AI Gets Wrong About Lab Inquiries

The opportunity to gain research experience during your undergraduate degree is a huge perk of attending a research-intensive university, but one which has long been buried in the hidden curriculum, accessed only by students lucky enough to have a mentor who understood the pathway. The good news is that efforts to increase accessibility have resulted in a steep increase in inquiries from students of varied backgrounds and perspectives; the bad news is that competition for a limited number of spots has grown alongside interest. Distinguishing yourself from the pack matters now more than ever.

This rapid growth in interest (and the influx of generic inquiries that had clearly been emailed en masse to the entire department) exposed another unwritten rule: how to actually get a researcher to notice your application. The advice we gave was straightforward - there are hundreds of students emailing dozens of researchers; if you want to stand out you need to make it clear why this program of research is of interest to you, and what you bring that makes you the right fit for the role. To their credit, students picked up the formula quickly: read a paper from the lab, comment on what you found interesting, note the attributes that suggest you'd thrive in a fast-paced research environment.

It makes sense, then, that AI picked up the same formula. It's a clear, well-defined pattern - the kind of thing a language model is good at reproducing. The problem is that it reproduces the template of a strong inquiry without the substance behind it, and the gap between the two is obvious the moment you've read more than one:

Dear Dr. [Name],

My name is [Name] and I am a student in the [program title] program at [institution]. I was highly impressed by your work in the area of [summary of research program from faculty website] — particularly your recent paper on [title of random paper pulled from Google Scholar]. I'm eager to build hands-on experience in the lab and would be happy to discuss ways I might contribute to this program of research. I am a [three positive attributes] student eager to make a meaningful contribution. I look forward to discussing the opportunity further.

Sincerely, 

[Name]

I understand the appeal. AI is very good at producing something that looks like a personalized email on the first read, and the shortcut is one simple prompt away, while the alternative is an afternoon of actual reading. However, what AI fails to deliver is the part that actually makes an email personal: a real opinion about the work, a sense of connection to the field of study, and a meaningful description of what you want out of the experience. As a result, these AI-generated inquiries read as generic in exactly the way the original, unpersonalized emails did a decade ago. Same problem, better disguise, and just as easy to spot.

So what actually works? The honest answer is that there isn't a shortcut here. A good application requires real work to communicate a genuine interest in a specific research program and signal the effort you plan to put into the work should you be offered a position. Here's how I think this can be accomplished:

1. Understand your "why" before you start emailing anyone. What about research interests you? What questions have stuck with you from a course, a job, or your own experience? This isn't a throwaway step - it's what will actually distinguish your email, because it's one part of the process AI can't do for you.

2. Let that narrow your list. I'm asking you to invest more time per application, so a shorter, better-targeted list serves you better than a long one. Five labs you're genuinely curious about will get you further than twenty you're not.

3. Read a couple (or more) recent papers properly, not just the abstract. You're looking for something specific enough to spark a discussion. This could be a method you're curious about, a finding that surprised you, or an area of potential follow up you noticed. That's the difference between "I was impressed by your recent paper" and an actual comment or question about the work.

4. Talk to someone already in the lab, if you can. Current students and trainees are usually happy to talk about what the day-to-day is actually like, and it'll help you speak to fit in a way a faculty bio never will.

That kind of groundwork changes the email completely:

Dear [Name],

My name is [Name] and I am a student in the [program title] program at [institution]. I recently finished a course in [topic A] that sparked a real interest in how it presents in [population B] - an interest sharpened by my own experience with [population B], which I think gives me a useful lens on the question. Your recent paper, [title], offered a framework I found genuinely useful, and it left me wondering [question]. I also reached out to [student name] in your group, and based on that conversation I think your mentorship style would be a strong match for the skills I'm hoping to develop, alongside my existing strengths in [relevant skills]. I'd welcome the chance to discuss any opportunities to get involved with the lab — please let me know if there's a time that would work for a brief conversation.

Sincerely, [Name]

To be clear, none of this is an argument against using AI. There are real places in this process where it can genuinely help (e.g., pressure-testing a research question, or figuring out which skills matter most for a given field). The issue isn't the tool, it's what you ask it to do. Ask it to help you think, and it can be a real asset. Ask it to do the thinking for you, and it will happily hand you something that looks finished but isn't. The problem is, that gap is visible to exactly the person you're trying to impress. There's no version of this where the shortcut works - the effort is the point.

tl;dr

1. Lab spots are competitive, and standing out matters more every year

2. AI learned from the early advice given to students about personalizing their inquiries, but their output remains a highly superficial representation

3. Real differentiation comes from the groundwork that AI cannot reproduce: know your own interests, narrow your list to labs that align with those interests, read the papers, talk to the students

4. AI is a good thinking partner and a bad substitute for thinking. There's no shortcut that skips the effort and still stands out