A new chatbot billboard recently appeared in San Francisco. It looks almost identical to the dozens of AI advertisements that have popped up throughout the city in recent years: simple font, a black-and-white color palette, and copy that sounds like it was written by a robot.
“ChatTJB.org,” the billboard reads. “The leading chat interface powered by AI.” An asterisk after that final word leads readers to the catch: “AI,” in this case, stands for “Actual Individual.”
This billboard wasn’t sponsored by a company, but by Tucker Bryant, a former marketing manager at Google who now spends his time as a keynote speaker, artist, poet, and social media creator. ChatTJB is his latest digital project, and it’s a very real chatbot service that funnels every prompt to Bryant himself, who responds to them to the best of his ability.
If Bryant’s project sounds familiar, that’s probably because other artists have been using similar tactics to express their dissent against companies peddling their artificial intelligence features for the past several months. As tech hubs like New York City and San Francisco become inundated with AI startups and advertising, independent artists are creating their own subgenre of digital and physical parodies that take aim at AI “slop” advertising, and ask the public to think critically about their relationship with AI.
“Hire everyone on earth without doing anything.” “The world’s first wearableAI grandfather.” “UnderwAIr: Panties that think.”
All of these are examples of ads that, like Bryant’s physical billboard, have popped up IRL across the New York City subway system. They’re the creations of two comedians, Harris Alterman and Dave Ross, who have spent the past two months making and displaying fake AI ads in the city.
In an early June interview with Fast Company, the pair explained that their aim is to skewer what they call “slop voice”: a distinctive advertising format that uses techy buzzwords, minimalist design, and zero-context claims to capture a high tech, SaaS-speaking in-group.
Slop voice, as Bryant points out in an August 2 Instagram video debuting his billboard, is a similarly rampant problem in San Francisco. Real examples he’s encountered include, “Goodbye missed calls. Hello AI receptionist”; “Agents that improve themselves start with evals”; and “No AI without APIs.”
“San Francisco has AI billboards absolutely everywhere, and it sort of felt like they popped up almost overnight,” Bryant says. “It can feel strange being in this city where on one hand, AI optimists make big promises and AI is doing some genuinely incredible things, but where it also has people anxious about the future. That cognitive dissonance can feel sort of surreal at times, so the billboards that are such a prominent visual feature of the city felt like a natural place to prod at AI’s cultural omniscience.”
Like Alterman and Ross, Bryant decided to counter these befuddling messages with one of his own. He paid $6,000 to rent a billboard in downtown San Francisco for the month of July. However, ChatTBJ takes the premise one step further by directing viewers to a working interface where they can chat with Bryant himself.
While Bryant’s billboard went up in early July, ChatTJB has been live since April. The interface, which Bryant says he designed on the AI-powered website builder Lovable, looks a lot like an average chatbot interface, including a nebulous void logo and simple conversational interface. Unlike an AI chatbot, though, ChatTJB might take a while to respond, and its answers come from one guy’s memory rather than a huge database. “Initially, I made it as a sort of informal commentary on cognitive surrender—the idea that people (myself included) sometimes fall into the trap of trusting LLMs [large language models] so readily that we stop thinking critically about the answers we’re receiving and grow dependent on them for things we’re capable of working through ourselves,” Bryant says.
When the site went up, Bryant was responding to an average of about 1,000 people per day about what color to paint their house, what to have for lunch, and how to handle awkward situations. After he bought the billboard and posted it to his Instagram, that influx soared to around 5,000 prompts every hour. Given the massive volume, Bryant has d the site while he works on vetting volunteers to answer inquiries and building safety infrastructure like automatic message filtering and crisis resources for people in distress.
Bryant’s crowdsourcing model for answering AI prompts echoes another interface that gained popularity earlier this year. Called “Your AI slop bores me,” the website took off in March after netizens discovered that its drawing and writing functions were fully powered by users. At the time, its programmer, Mihir Maroju, told Fast Company that inspiration for the site came from “a frustration for AI art and its proliferation, making artists’ lives worse and also just filling the internet with low-effort generic slop.”
Other creators have used more direct strategies to help AI users think critically about their relationship with chatbots. In December 2025, the artist Tega Brain released Slop Evader, a Chrome extension built to sift AI slop from search results. And this March, the designer Sam Lavigne debuted Slow LLM, an add-on built to make AI chatbots respond at a torturously slow pace.
As many AI executives and startup advertisements peddle the narrative that universal AI adoption is inevitable, artists are finding their own subtle ways to poke holes in that argument—and encourage others to do the same.
“The hope was that it would lightly nudge people to and remember, ‘Hey, even though this experience looks and feels futuristic, maybe LLMs don’t always actually have the absolute best answer to everything and my brain still has a purpose here,’” Bryant says.