I Chatted With Google’s Lifesize, Hyperreal AI Companion
Sitting in a comfortably air-conditioned cabin in the superheated outdoor demonstration landscape on Google I/O This week I had a nice and strange conversation with a smiling person who wasn’t actually there. They weren’t even a real human being.
The stand was a demonstration installation of Google Beama telepresence video display device with a large screen camera. As we spoke face to face, it all seemed incredibly real.
I’ve experienced demonstrations of Google Beam before, back when it was called Project Starline: with real humans on the other end, in holographic 3D without glasses. It was a whole new twist, and even though the demo was in 2D, it could easily embrace 3D rendering as well.
@scottsteinsayshi A strange and hyper-realistic AI video agent demo via Google Beam on Google IO. And yes, it created AI images of me. Where the hell will this end up and will he aim to replace a co-worker…or end up in a hotel or theme park?
♬ original sound – scottstein89
Google plans a wider rollout later this year for Beam, its business-focused collaborative video chat technology created with HP. The central idea of Beam is to connect two people remotely as if they were sitting around a table in person. This requires two people with Beams though. Google’s next step is to imagine situations where other people are more easily brought in from other devices, or where there are no other people, period.
Watch this: Is Google’s strange virtual human a future colleague or janitor?
The bespoke AI video agent, created with internal models that Google hasn’t shared, actually shocked me. Like a real deepfake, the woman-like agent was photorealistic (the agent didn’t have a name, I just started talking with him), and smiled, gestured, and talked to me casually. This agent was there just to help me and chat casually, just like Gemini or any AI chatbot.
I asked him to generate a photo of me doing magic tricks at a New York Jets game, and he happily obliged. He asked me about the bananas on the table and complimented my cameraman’s backpack. It gave us map search recommendations from a contained demo experience provided by a Google employee while I watched, as he gestured to the maps and images next to me.
Yes, I still perform magic at a New York Jets game – it’s my Will Smith pasta AI test.
It was strange. This video agent, even though he wasn’t in 3D (just standard 2D), was one of the most real people I’ve ever chatted with before.
But do we need Beam? Many of us already have telepresence at our fingertips, when video chatting on our phones. For those who want it, there is connection via VR headsets, or future iterations of telepresence, as they may emerge. on AR glasses. Microsoft just stopped trying to evolve its telepresence realism in Teams. Will Google Beam make enough of a difference for businesses and institutions to convey a sense of real presence on a larger scale? And would an AI agent video assistant make things helpful or annoying? And what jobs could a video agent like this replace?
According to Andrew Nartker, general manager of Google Beam, who walked me through demos, this video AI agent is largely an experiment. But it’s also something that could land in an office as well as in a public place, like an interactive talking kiosk.
On the whimsical side, the demo made me think of theme parks. Would a 3D light field version of an AI character greet me at a food stand in a future Disney Star Wars Galaxy’s Edge expansion, serving me interplanetary food while wrinkling its alien face? Would a themed hotel use this to create a magical experience? Or would it be used to replace real janitors and workers in certain situations, a more advanced concept of the walk-through windows of AI chatbots that already exist?
I can see all these ways at once. The Google Beam demo, however, impressed and blew me away – and concerned me – more than almost anything else at this year’s Google I/O.
