In the not-too-distant future, Zoom users could send AI avatars to attend meetings in their absence, the company’s CEO has suggested, delegating the drudgery of corporate life to a system trained on their own content .
Such a system would be “five or six years” away, Eric Yuan told The Verge magazinebut he added that the company is working on nearer-term technologies that could bring it closer to reality.
“Let’s assume, fast forward five or six years, that AI is ready,” Yuan said. “AI can probably help with maybe 90% of the work, but in terms of real-time interaction, today you and I are talking online. So, I can send my digital version, you can send your digital version.”
Using AI avatars in this way could free up time for less career-oriented choices, added Yuan, who also founded Zoom. “You and I may have more time for more personal interactions, but maybe not for work. Maybe for something else. Why do we have to work five days a week? On the road, four days or three days. Why not spend more time with your family?”
Eventually, he suggests, each user will have their own “large language model” (LLM), the underlying technology of services like ChatGPT, which will be trained on their own speech and behavior patterns, so they can generate highly personalized responses to queries and requests.
Such systems could be a natural progression from AI tools that already exist today. Services like Gmail can summarize and suggest responses to emails based on previous messages, while Microsoft Teams will transcribe and summarize video conferences, automatically generating a to-do list from the content.
Other services will generate realistic video avatars and plausible generated speech from a text transcript. Put them all together, and it can feel like an AI avatar is tantalizingly close.
However, AI expert Simon Willison dismissed the idea that such technology is imminent or even possible. “My fundamental problem with this whole idea is that it represents pure AI science fiction thinking,” he said. “Just because an LLM can give an acceptable impression of someone, it does not mean that he can actually perform useful ‘work’ on that person’s behalf.
“LLMs are useful tools for thinking. They are terrible tools to delegate decision making to. This is currently my red line for using it: anytime someone outsources real decision-making authority to an opaque random number generator is a recipe for disaster.”
Others expressed concern about the blurring of real and fake. Steve Won, the chief product officer of security and identity company 1Password, pointed to Yuan’s claims as proof that online authentication is about to get significantly more complicated.
to newsletter promotion
“How many digital twins can I have at any given time? It’s like a Max Headroom situation,” Won said Tuesday, referring to the 80s television series.
“The fact that the leading virtual communication app in the world thinks, ‘yes, it’s perfectly fine to have fake conversations, represent and make business decisions’ – I think that makes it a burning problem that we’re going to have to solve.” “
