AI Resume Parser Reads Grocery List, Flags 'Milk' as Key Skill

A recruiting coordinator at a mid-sized tech company discovered their new AI-powered resume parser had been analyzing the wrong document for three weeks. The system had been screening a candidate's grocery list instead of their actual resume, and somehow the candidate made it to the phone screen stage.

The Mix-Up

The candidate had apparently attached "shopping-list-final.pdf" instead of "resume-final.pdf" when applying for a senior software engineering position. The AI parser, confident in its abilities, proceeded to extract and analyze the contents with zero hesitation.

The system flagged several "key competencies" including:

• Milk (listed as "dairy management experience")
• Eggs (interpreted as "protein optimization")
• Bread (categorized under "staple goods procurement")
• Coffee (flagged as "productivity enhancement materials")

The algorithm gave the candidate an 87% match for the role. The grocery list mentioned "whole wheat," which the AI interpreted as experience with "full stack" development. "Organic" was read as expertise in "organic growth strategies."

How It Got This Far

The coordinator who initially reviewed the application saw the high match score and the AI's confident assessment of "strong foundational skills" and moved the candidate forward. Because why would you question the robot?

The hiring manager, drowning in their inbox, saw "dairy management" and assumed it was a typo for "data management." They scheduled a phone screen.

The candidate, realizing the mistake approximately 30 seconds after hitting submit, figured they'd get an instant rejection and didn't bother following up. When they got a phone screen invite two days later, they were confused but not confused enough to turn down a potential job.

The Phone Screen

The recruiter opened with "I see you have extensive experience with optimization and procurement strategies." The candidate, thinking quickly, just started talking about their actual software engineering background. The recruiter, who hadn't actually read the grocery list because they trusted the AI's summary, thought everything was going great.

It wasn't until the hiring manager prep meeting that someone pulled up the actual "resume" and saw line items like "2% milk - the regular one not the fancy organic" and "bananas (not too ripe)."

The Aftermath

The company is now adding a "sanity check" feature to flag documents under 200 words or containing certain keywords like "checkout" and "aisle." They're also making humans look at the actual document before phone screens, which feels like something they should have been doing anyway.

The candidate did get a proper interview after submitting their actual resume. The hiring manager appreciated their "ability to think on their feet" during that first phone screen. They're currently in the final round.

The AI parser has been demoted to "assistant" level access and is no longer allowed to make independent decisions. It's basically doing the same job as before, but now someone checks its work.

The grocery list, meanwhile, has been printed out and hung in the recruiting team's break room as a reminder that AI is only as smart as the humans who trust it blindly.

The candidate's grocery list also included "ice cream (I deserve it)," which the AI flagged as "self-reward compensation philosophy." That part was actually accurate.