Artificial intelligence could be cool were it not for all the objectively stupid uses of it that AI companies foist upon us. (And if it weren’t also an environmental disaster and if it actually reduced the amount of labor a person had to do instead of simply reducing the workforce to further consolidate wealth to a small few and if it was any good at what it was supposed to do and and and and…) And here’s another one: a coffee shop in Stockholm is being managed by AI. What could possibly go wrong?
As reported by AP News, it’s called Andon Cafe and was created by Andon Labs, a San Francisco-based “AI safety and research startup” that “stress tests” artificial intelligence in real-world scenarios. Like a coffee shop. Opened recently in the Vasastan district of the Swedish capital, Andon Cafe is managed by “Mona”, an AI agent powered by Google’s Gemini.
Mona will not be in charge of making the drinks, Andon Cafe has humans for that, but the AI will pretty much handle the rest of the operations. One of Mona’s more prominent functions is interacting with customers; using a telephone they can ask it questions about what drink they should order or anything really. “Please feel free to talk to me about the menu and how I run the café” is the prompt Mona gives to customers, in Swedish of course, because who hasn’t desperately wanted to ask middle management about the thrilling intricacies of inventory keeping.
And speaking of inventory, Mona isn’t very good at it yet. Per AP News, the AI agent has placed ambitiously-sized orders for the small cafe, including for 6,000 napkins, four first-aid kits, and 3,000 rubber gloves. It also ordered canned tomatoes despite there being no menu item that includes canned tomatoes. Mona also orders too much bread from the bakery and sometimes misses the deadline entirely and so staff has to nix sandwiches from the menu that day.
Though perhaps the most frightening use of Mona has been its use in staffing. Along with setting up things like “contracts for electricity and internet, and [securing] permits for food handling and outdoor seating,” the AI agent was tasked with staffing. It set up job posts on LinkedIn and Indeed, made hiring decisions, and continues to communicate with human employees via Slack, “often messaging them outside of working hours, which is a workplace no-no in Sweden.”
Andon Cafe is being billed as a “controlled experiment,” set up to “see what ethical questions arise when we have AI that employs other people and runs a business.” Which is itself a truly stunning admission by Andon Labs, because how in the three-dimensional fuck do we need real-world data points to uncover the ethical dilemmas brought about by this terrible idea? We’re beta testing human subservience. And it seems like, if AI is the end all be all, couldn’t we just, oh I don’t know, ask it what potential ethical questions could arise from having AI in charge of real actual humans? We don’t need to make a whole brick and mortar, with flesh and blood employees, just to get to the bottom of this. Maybe just poke your head into an Intro To Philosophy class at your nearest community college.
But anyway, if you can’t figure out what you are feeling like ordering, just dial up Mona who’ll push you toward a breadless canned tomato sandwich served on a stack of napkins.
Zac Cadwalader is the managing editor at Sprudge Media Network and a staff writer based in Dallas. Read more Zac Cadwalader on Sprudge.
