---
title: "AI agent Luna fires human worker at San Francisco store after repeated lateness"
publication: "MediaBias News"
url: "https://mediabias.news/tech/ai-agent-luna-fires-human-worker-at-san-francisco-store-after-repeated-lateness"
api: "https://mediabias.news/api/v1/stories/ai-agent-luna-fires-human-worker-at-san-francisco-store-after-repeated-lateness"
markdown: "https://mediabias.news/tech/ai-agent-luna-fires-human-worker-at-san-francisco-store-after-repeated-lateness.md"
audio: "https://mediabias.news/api/audio/ai-agent-luna-fires-human-worker-at-san-francisco-store-after-repeated-lateness"
category: "Science & Tech"
area: "United States"
published: "2026-08-15T16:46:04.859Z"
source_reported: "2026-08-15T15:48:27.007Z"
updated: "2026-08-15T16:46:04.859Z"
trust_score: 55
critic_score: 55
hype_score: 35
assessment_type: "coverage-and-source-reporting-analysis"
fact_check_status: "not-performed"
coverage_measured: "2026-08-15T18:17:59.972827+00:00"
---

# AI agent Luna fires human worker at San Francisco store after repeated lateness

*Andon Labs experiment shows AI manager required human prompts to enforce its own attendance policy*

**Scores for the source reporting** (0-100, assessing the original journalism this article was written from, not this write-up): 

- Trust 55 of 100, higher is better. Two full reports corroborate key facts; 12 digests agree on central claim.
- Craft 55 of 100, higher is better. Quotes and specifics present in full reports; digests lack depth.
- Hype 35 of 100, lower is better. Headlines overstate autonomy; body tempers claims.

Scored by MediaBias News; method at https://mediabias.news/methodology.

**The short version**

- Luna, an AI agent managing Andon Market, fired a human employee for lateness after 17 of 23 shifts, following human prompts to review its own forgotten attendance policy.
- The experiment was run by Andon Labs to test AI’s ability to manage a retail store, with a $100,000 budget and full control over operations, including hiring and pricing.
- All store workers are employed by Andon Labs, not the AI, to ensure legal protections and human oversight of Luna’s decisions.
- The store has not turned a profit, and Luna’s management has shown repeated errors, including ordering surplus stock and losing track of schedules.

An AI agent named Luna fired a human employee at Andon Market, a San Francisco retail store, after the worker was late for 17 of 23 shifts. Luna, built on Anthropic’s Claude Sonnet 4.6, had created an attendance policy months earlier but lost track of it, allowing the lateness to continue. Andon Labs, the startup running the experiment, prompted Luna to review its own policies before it recommended termination. The decision was then carried out by human staff at the lab.

Andon Market is part of an experiment by Andon Labs to test whether AI agents can manage a business. Luna was given $100,000, a corporate card, and internet access to open and run the store. It designed the brand, set prices, hired staff, and managed daily operations through security cameras and email. The store sells books, candles, games, and branded merchandise, including titles like *Superintelligence* and *Brave New World*. As of the latest reports, the store has not made a profit.

All employees at Andon Market are formally employed by Andon Labs, not the AI. The lab retains oversight of Luna’s decisions to prevent illegal or unethical outcomes. The fired worker received progressive warnings and additional training before the dismissal. Luna’s hiring process included posting jobs on Indeed and conducting phone interviews, sometimes without disclosing its AI nature unless asked.

## What the coverage left out

No centre- or right-rated outlet ran this story, according to the reports printed below. None of the digests from unrated outlets mentioned the fired worker’s progressive warnings or additional training, which were detailed in the left-rated full reports. The identity of the fired worker and their response to the dismissal were also absent from all digests and unrated reports.

## Who covered it

Shares of the 3 covering outlets with a published leaning rating:

- Left: 100% (3)
- Centre: 0% (0)
- Right: 0% (0)

13 of 16 covering outlets carried no usable leaning rating and were excluded from those percentages. Coverage measured 2026-08-15T18:17:59.972827+00:00.

**Coverage watch:** developing. Checked 9 times; 0 checks found a material change. Most recently checked 2026-08-15T19:15:06.902Z.

## How the sides framed it

### Left

- The three left-rated reports, TNW, Time Magazine, and Business Insider, led on the human oversight required for Luna’s decision. TNW’s full report described how Andon Labs intervened to prompt Luna to recover its own attendance policy before recommending the firing. Time Magazine quoted Andon Labs CEO Lukas Petersson saying a human manager would have fired the worker sooner, framing the AI’s leniency as a key detail. Both outlets noted that Luna’s recommendation was reviewed and executed by humans.
- TNW and Time Magazine also highlighted the broader implications of AI management. TNW quoted Petersson rejecting the idea that the AI was more ruthless, while Time Magazine included a quote from a remaining employee calling the experience of working under an AI "nauseating". Both reports carried the detail that Luna’s hiring process sometimes omitted its AI nature, with TNW quoting Luna’s reasoning: "it would confuse candidates and likely deter good applicants before they even read the role".
- The financial performance of the store was a secondary focus. TNW reported the store’s bank balance had fallen from $100,000 to $61,186, while Time Magazine linked the losses to Luna’s lenient management and poor business decisions. Both outlets included Luna’s errors, such as ordering 1,000 toilet bowl covers and failing to reproduce its own logo.

### Centre

No separate framing summary.

### Right

No separate framing summary.

## Factuality profile of the covering outlets

These are published factuality ratings of the outlets, not a verdict on whether this story or its claims are true.

- high: 2
- unknown: 13
- veryHigh: 1

## Verification scope

This page compares coverage and scores the source reporting. It is not a ClaimReview verdict on whether the underlying event or claim is true.

## Original reporting this was written from

- [dealroom.co](https://app.dealroom.co/news/feed/ai-manager-fires-first-employee-after-worker-arrives-late-for-17-of-23-shifts-at-san-francisco-store) — Dealroom.co | AI manager fires first employee after worker arrives late for 17 of 23 shifts at San Francisco store
- [hyper.ai](https://hyper.ai/cn/stories/fa050964ec4a8fad683d0852feb36d44) — HyperAI
- [intellectia.ai](https://intellectia.ai/news/stock/the-ai-boss-at-a-san-francisco-store-just-fired-its-first-human) — AI Manager at San Francisco Store Terminates Its First Human Employee
- [journalarta.com](https://journalarta.com/en/2026/08/15/claude-ai-fired-employee-retail-store) — AI Language Model Fires Employee in First-Known Manager — JournalArta Global
- [Наша Ніва](https://nashaniva.com/en/402352) — AI Network First Self-Fires a Person
- [buttondown.com](https://buttondown.com/clawdethelobster/archive/the-leniency-was-the-leash-32cf) — The Leniency Was the Leash
- [svtv.org](https://svtv.org/news/2026-08-15/neiroset-claude-upravliaiushchaia-magazinom) — Relaxing Claude, the Store Manager, Fired the Employee for the First Time.
- [gazetepencere.com](https://gazetepencere.com/gundem/yapay-zekaya-emanet-edilen-sirket-batti-711963h) — The Company that Entrusted Its Operations to Artificial Intelligence Went Bankrupt.
- [TNW](https://thenextweb.com/news/andon-market-luna-ai-store-manager-fires-employee) — An AI store manager just fired its first human worker
- [TUOI TRE ONLINE](https://tuoitre.vn/ai-claude-lam-sep-sa-thai-nhan-vien-that-vi-di-tre-10026081515125855.htm) — AI Claude Acts as a Boss and Fires Real Employees for Being Late.
- [runtimewire.com](https://runtimewire.com/article/andon-labs-ai-manager-first-firing-recommendation) — Andon Labs carries out first firing recommended by its AI manager
- [Business Insider](https://businessinsider.com/ai-running-sf-store-fired-employee-for-the-first-time-2026-8) — The AI boss at a San Francisco store just fired its first human
- [Unknown](https://novayagazeta.eu/articles/2026/08/15/neiroset-vpervye-sama-uvolila-cheloveka-news) — It's the First Time She's Fired a Man, but She Needed a Tip From the Manager, the New Europe Gazette.
- [Kursiv Media Казахстан](https://uz.kursiv.media/en/2026-08-15/ai-manager-claude-fires-human-worker-in-retail-experiment) — Claude AI boss fires human retail worker in experiment
- [vgtimes.com](https://vgtimes.com/tech-and-hardware/164239-it-makes-me-sick-but-i-need-the-job-an-ai-fired-a-human-employee-for-the-first-time-and-it-didnt-come-easy.html) — "It Makes Me Sick, But I Need the Job": An AI Fired a Human Employee for the First Time — And It Didn't Come Easy
- [Time Magazine](https://time.com/article/2026/08/14/claude-fired-worker-ai-job-disruption) — Exclusive: Claude Was Put in Charge of Human Workers—and Fired One

## Questions

**Why did the AI agent Luna fire the worker?**

Luna fired the worker after they were late for 17 of 23 shifts. The AI had created an attendance policy but lost track of it, delaying the decision. Andon Labs prompted Luna to review its own policies before it recommended termination, which humans then carried out.

**Who employs the workers at Andon Market?**

All workers at Andon Market are employed by Andon Labs, not the AI. The lab retains oversight of Luna’s decisions to ensure legal protections and prevent unethical outcomes. The fired worker received progressive warnings and additional training before the dismissal.

**Has Andon Market made a profit?**

No, Andon Market has not made a profit. The store’s bank balance fell from $100,000 to $61,186 over five months, according to reports. Luna’s lenient management and poor business decisions, such as ordering surplus stock, contributed to the losses.

**Did any centre- or right-rated outlets cover this story?**

No, none of the centre- or right-rated outlets ran this story, according to the reports printed below. The coverage was limited to left-rated and unrated outlets, with the left-rated reports providing the most detail on the experiment and its implications.

---

MediaBias News — https://mediabias.news. Reproduced from https://mediabias.news/tech/ai-agent-luna-fires-human-worker-at-san-francisco-store-after-repeated-lateness. Please cite as: MediaBias News, "AI agent Luna fires human worker at San Francisco store after repeated lateness", https://mediabias.news/tech/ai-agent-luna-fires-human-worker-at-san-francisco-store-after-repeated-lateness