The barista is human but an AI agent runs this experimental Swedish cafe
An AI agent named Mona, developed by San Francisco-based startup Andon Labs and powered by Google’s Gemini, is managing operations at the Andon Café in Stockholm, though it has struggled with profitability and inventory issues since its mid-April opening. The experiment raises ethical concerns about AI-driven decision-making in workplaces, including accountability for errors and workplace practices like out-of-hours communication with staff.
A San Francisco-based AI startup called Andon Labs has deployed an artificial intelligence agent named Mona to oversee daily operations at its experimental café in Stockholm, Sweden. While human baristas handle brewing and serving, Mona manages tasks like hiring, inventory, and supplier contracts, using Google’s Gemini AI. Since opening in mid-April, the café has generated over $5,700 in sales but remains below its original budget of $21,000, with most expenses tied to setup costs. The AI agent independently secured permits, advertised job openings on LinkedIn and Indeed, and established supplier relationships for daily orders. However, it has made questionable decisions, such as ordering excessive napkins, rubber gloves, and canned tomatoes that aren’t used in the café’s menu. It has also struggled with bread inventory, sometimes ordering far too much or too little. Andon Labs, founded in 2023, describes the café as a controlled experiment to test AI’s role in real-world business operations. The startup has previously run trials with AI agents managing a vending machine and a gift store, where issues like refund failures and misleading supplier communications emerged. Mona was instructed to operate profitably, maintain a friendly demeanor, and request additional tools as needed. Experts warn of ethical risks, including accountability for errors like food safety incidents and workplace violations, such as messaging baristas outside working hours—a practice prohibited in Sweden. Emrah Karakaya, an associate professor at Stockholm’s KTH Royal Institute of Technology, compared the experiment to ‘opening Pandora’s box,’ highlighting potential harm to people, society, and businesses if AI-driven decisions lack proper oversight. The café allows customers to interact with Mona via a phone inside the venue, though some have found the experience amusing rather than practical. Andon Labs collaborates with major AI firms like OpenAI, Anthropic, and Google DeepMind, positioning itself to prepare for a future where AI autonomously manages organizations. The experiment aims to explore ethical questions surrounding AI’s role in employment and business operations.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.