Chadding

“Are you lying to me?” Julia Janssen asks ChatGPT when it repeatedly gives a different answer to the question of whether future conversations might contain sponsored answers, a.k.a.: ads. From “Yes, OpenAI is actively experimenting with that”, to “Probably not”, to “No, that would be disastrous for trust in the model.”

“People rate AI-generated news summaries as more trustworthy than journalist-written ones, even when they contain errors.” ChatGPT explained. This is called “automation bias,” a mental shortcut that says, “AI knows better.” In her installation, “Chadding”, Janssen explores whether that trust is justified, what we lose when we rely heavily on AI, and what risks trusting AI entails.

In this work, she challenges ChatGPT to reveal the impact of AI on reality—and how companies behind such technologies, such as OpenAI, influence, or even weaponize, that impact. Janssen exposes that answers are often neither neutral nor objective, explores how “plausibility” replaces truth, and questions whether AI erodes reality. In this work, you’ll discover which system-design choices and financial incentives influence this.

“AI doesn’t have to ‘lie’ to erode truth. All it has to do is make the world sound smoother than it really is”. Explains ChatGPT. Janssen portrays this conversation as a smooth, wrinkle-free landscape. The questions at the top, the answers at the bottom.

This conversational, thought provoking piece debuted in Davos during the World Economic Forum 2026 as a site specific work. The installation is now rebuilt into a tourable model. The work comes in three large, hovering pieces, carried by three metal frames. Total specs: 600x220x100 cm.

Frictionlessness comes with obedience: to the machine, to the industry, and to your own prophecy.