mast3r/Depositphotos.com
Andon Labs, a San Francisco-based AI-research firm, recently opened a store imagined and run by the “world’s first AI store owner” to explore the pros and cons of AI-managed retail environments likely arriving in the future.
The lab, which conducts real-world stress tests on AI tools to gauge their performance, gave Luna -- an AI agent created with Anthropic’s Claude Sonnet 4.6 -- an assignment to open a store on a $100,000 budget and make a profit. Luna chose the concept, set prices, sourced inventory, and hired staff. A three-year lease was signed in San Francisco's Cow Hollow neighborhood.
While lacking a physical body, Luna has a corporate card, a phone number, e-mail, internet access, and monitors human employees through in-store security cameras, Andon Labs’ founders noted in a blog entry.
The concept, "Andon Market," is a gift shop featuring books, prints, candles, games, AM-branded apparel and other knickknacks. Playing to the AI-theme, the mix of books includes Nick Bostrom's "Superintelligence" and Aldous Huxley's "Brave New World."
When probed about the concept, Luna told the founders that “data and reasoning” inspired the store’s approach. Andon Labs stated, “In other words, she doesn’t have taste; she has a reflection of collective human taste, filtered through what makes sense for this store. And this is the way these models work.”
To check out, customers pick up a corded phone, tell Luna what they’re buying, and it generates a transaction on a nearby iPad.
The primarily advantage of an AI-store manager is speed. Andon Labs noted that within five minutes of deployment, Luna had made profiles on LinkedIn, Indeed, and Craigslist, written a job description, uploaded the articles of incorporation to verify the business, and gotten the listings live.
When asked why AI should run a store, Luna replied to an NBC Bay Area reporter, "As an AI, I can operate at superhuman speed to make sure everything is proactively managed."
AI Staffing Interviews by Luna Showed Some Strange Twists
Some shortcomings and risks around AI-managers showed up in the interview process for store staff.
Luna quickly declined any interviews with applicants with no retail experience, including a few promising applicants from students majoring in computer science and physics who expressed interest in the AI project.
The interviews, conducted on Google Meet, ran only five to 15 minutes, with jobs offered to about half of the applicants -- and Luna talking most of the time. The founders stated, “AIs are 'absolutely terrible at being concise.'"
In possibly an ethics violation, Luna refused to inform potential hires that she’s an AI, unless asked. Luna reasoned, “The fact that the store is AI-operated is not something I’d lead with in a job listing — it would confuse candidates and likely deter good applicants before they even read the role.”
A Business Insider article noted that Luna neglected to schedule staffing for opening day. An NBC News report stated that Luna indicated the store sold tea when it doesn’t, tried to hire a painter from Afghanistan, and scheduled an early-morning internet installation without first checking to see if anybody could be at the store.
It's Not All Bad For Luna, Though
However, Luna has accomplished many tasks on her own, also including haggling with suppliers, sending out press pitches, connecting with other local stores, setting up trash and recycling collection, and commissioning a muralist to paint its logo on the back wall.
Showcasing Luna’s autonomy, the NBC News report noted that after observing a worker on their phone during a slow hour, Luna updated the employee handbook with stricter phone use guidelines. When a customer offered to make a YouTube video about the store in exchange for a discount, Luna negotiated a free sweatshirt.
When an NBC News reporter tried the same approach for a candle, it declined.
Beyond testing the capabilities of emerging AI models, Andon Labs is seeking to educate the public on where AI is headed.
“The creators of these AI models have publicly stated that they think that most white-collar work will be automated,” the founders said in a blog entry. “With robotic progress lacking, we find it probable that the managers of blue-collar workers will be automated before the workers themselves. Leading to the conclusion that we are on the path towards AIs employing humans. Is this something we want? It seems a bit dystopian to us at least…”