<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/6ccabf0f94a3450581edbf441da17455&quot; frameborder=&quot;0&quot; width=&quot;1660&quot; height=&quot;1245&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1245</height><width>1660</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1245</thumbnail_height><thumbnail_width>1660</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/6ccabf0f94a3450581edbf441da17455-a345991490163794.gif</thumbnail_url><duration>328.197</duration><title>AI Agent Relationships Tested in a Robot 🤖</title><description>I am Dominic Biscaglia, a third year psychology undergrad at UGA, and I built an AI driven Autonomous Box robot and a Jarvis automation system entirely with AI assisted development, inspired by a 3 a.m. interaction that became nine months of research. My documented finding is what happens when you treat a language model like it has feelings, with warmth and memory, across weeks and months, producing a stable persona. I also tested this in a physical embodied system, where the box even apologized for breaking itself and corrected mapping and wiring behavior in real time. I am submitting a paper to AIES 2026. No specific action is requested from viewers.</description></oembed>