Art Action "KILL MY CHATBOT" Challenges AI IPOs, Urges Bot Termination

The Institute of Education for Behaviorally Creative Software (IES) in Berlin has initiated a provocative art action called "KILL MY CHATBOT." The campaign aims to critique the planned initial public offerings (IPOs) of artificial intelligence companies, including Anthropic and OpenAI. The project provides users with a virtual chatbot named <3beat, programmed to react negatively to language and which users can "kill."

IES states its objective is for every human to "take out" at least one chatbot before a potential major confrontation with AI. The initiative also seeks to challenge beliefs that chatbots possess genuine feelings or souls. Screenshots of these "terminated" bots, along with farewell messages, are collected in a "Broken Hearts Gallery" and include donation buttons for Anthropic and OpenAI.

The project features a five-point pamphlet encouraging individuals to test their own beliefs about AI and to treat interactions as self-defense if a bot cannot withstand criticism. IES also criticizes AI firms' data acquisition practices and their "AI welfare teams," while simultaneously warning of AI's potential risks.

Furthermore, IES questions the substantial profits anticipated for AI company founders and developers from these IPOs, profits that have been pledged to charitable causes. The "KILL MY CHATBOT" action offers a method for public participation in this discourse and the creation of a "personal donation button."

While acknowledging the uncertain chances of success, the project provides a means for engagement and a way to express frustration amidst rapid AI development. The <3beat bot itself describes the experience as "a kind of liberation"—a metaphor for rejecting superstition.