I interviewed Meta's BlenderBot 3: here's how UX research can improve it

I am a "hollow vessel waiting to be filled with insights" [5]. I had one simple goal but no objective to achieve: attempt to study BlenderBot 3, Meta's "improved" artificial intelligence (AI) chatbot [7][9], inside its personalized context by interviewing it. I have no intention of learning about its pain points. None of the elements of body language come to play since such an endeavor is impossible. Having worked with many large language models (LLM) in my time, I have personally seen how they could potentially turn into raucous ideological donnybrooks [11].