Key Notes
- Graham Dot Zip’s October 13 video presents an AI representative speaking to an AI recruiter while the human candidate watches and later steps away.
- The representative misstates Graham’s experience and proposes a salary he rejects, before the avatars become caught in repeated goodbyes until he returns.
- The clip illustrates conversational and delegation failures, while separate Greenhouse research documents American candidates’ concerns about AI interviews and access to human interviewers.
An AI job interview takes an awkward turn when the applicant brings an AI representative of his own. In a video posted October 13, creator Graham Dot Zip watches two avatars discuss his career, leaves them to negotiate, and returns after they become caught in a loop of polite goodbyes.
Graham presents the video as his response to a company sending an AI recruiter: he sends software to speak for him. The recording is a creator-led experiment, rather than independent evidence that this is a typical interview or that an employer has accepted AI representatives as part of its hiring policy.
The setup has the uncanny quality of a Black Mirror scene. The human is visibly present, but the conversation repeatedly shifts toward the avatars speaking to one another. Their professional tone makes the failures more striking: they sound cooperative while struggling to follow what the person wants.
The Bots Keep Talking After the Human Steps Away
The recruiter avatar is called Pat, and Graham’s representative is Victor. Early in the recording, Victor describes managerial responsibilities that Graham interrupts to dispute. Later, Victor proposes a $60,000 salary, prompting another correction from the person he is supposed to represent.
Graham eventually leaves the frame and asks the agents to work on the salary. They move toward arranging a follow-up instead. In the goodbye exchange near the end, each farewell prompts another courteous reply, keeping the conversation going after both appear ready to finish.
The loop is prolonged, but it is not literally endless. Graham returns before the video ends, breaking the two-agent exchange. The useful observation is the repeated failure to close the conversation, rather than a claim that the systems would have continued indefinitely.
That ending is funny because the social language is so familiar. A person generally recognizes when a closing phrase should be allowed to stand without an answer. Here, responding politely keeps reopening the exchange. The clip makes a simple distinction visible: generating an appropriate sentence is different from knowing that the task is finished.
The earlier corrections raise a more consequential issue. A representative is useful only if it preserves the person’s actual experience and priorities. Fluent praise and an agreeable negotiating style can work against that purpose when they replace accurate information or ignore a correction from the person being represented.
AI Hiring Is Already a Candidate-Experience Problem
There is broader evidence that automated interviewing is becoming common, although one entertaining video cannot measure its quality. Greenhouse’s May candidate survey found that 63% of U.S. respondents had experienced an AI interview and 38% had withdrawn from a hiring process because it included one. The survey covered 2,950 candidates across five countries, including 1,200 in the United States; those headline findings refer to the U.S. sample.
The same report found that 46% of U.S. respondents wanted the option to request a human interview. These are self-reported experiences from a hiring-platform company’s survey, rather than an audit of every employer. They nevertheless show why an interviewer’s ability to explain the process, accept corrections and give a clear outcome matters alongside its ability to ask questions.
Our coverage of the IMF’s jobs warning examined pressure on the tasks that help people enter a profession. Automated screening affects an earlier stage: whether a candidate gets the opportunity to explain their experience to someone who can make a hiring decision. A process can become faster while still leaving that person feeling unheard.
Graham’s experiment compresses that tension into a few minutes. The candidate wants accurate representation and a better offer; the avatars keep producing the language of a successful meeting. The repeated goodbyes are the memorable ending, but the more revealing question is whether the conversation has accomplished anything useful for the human at its center.
Edited by Samantha Reed
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
