Zuckerberg and Huang: I'm still not convinced by their vision for AI
When you buy through links on our articles, Future and its syndication partners may earn a commission.
Meta CEO Mark Zuckerberg and Nvidia CEO Jensen Huang swapped jackets and sat down for a fireside chat at SIGGRAPH 2024 on Monday, sharing their vision for the future of AI. It emerged that they have a lot in common, both imagining a future where every company, and even individual social media creators, will have an AI assistant or chatbot.
But while some of that vision sounds beneficial, or at least inevitable, I'm not entirely convinced that it's something most people want or need.
Zuckerberg used his appearance at SIGGRAPH to launch Meta's AI Studio, a platform for creating 'AI characters' – basically glorified chatbots with faces and "personalities". “Every single restaurant, every single website will probably, in the future, have these AIs,” Huang reckoned. Zuckerberg agreed, obviously. “Just like every business has an email address and a website and a social media account, I think, in the future, every business is going to have an AI,” he said.
Nvidia showed how much it's into the idea of digital avatars on Day 3 of SIGGRAPH with the unveiling of James, its digital and interactive brand ambassador. We only got a very brief glimpse: a 4-second video featuring one stilted line of dialogue. But we're told that by using Nvidia Maxine tech and the Nvidia ACE suite of generative AI tools, James will be able to connect with people emotionally and will be able to use humour. He'll also have fairly convincing facial expressions thanks to Maxine 3D and Audio2Face-2D.
But this is Nvidia's sector. Will your local Pizza delivery store want an AI avatar? Let's just recall that Meta's CEO thought we would all be living in the metaverse by now. He also thought people would want to buy designer outfits for their Instagram avatars.
In comparison with the metaverse call, Meta's AI chatbots, and even Nvidia's James, are a less radical conceptual leap for both companies and individuals. We're already used to this kind of frustrating interaction since so many companies are already using regular chatbots. Businesses will be keen to reduce human communication further if it means saving time and money. Even Instagram influencers may welcome no longer having to respond to their own DMs. But it could be risky.
Meta's billing its Creator AIs as "an extension of you", claiming that they will mimic users' tones and expressions, although responses will be labelled so users don't think they're speaking to a real person. But while users provide example responses to give an idea of how they want their bot to talk, the bots won't be trained on the sum total of their creators' knowledge but by a more generic AI model. That makes it likely that they will be able to spout facts (or falsehoods) and opinions that their creators may not know or agree with.
There are big questions around trust and credibility if companies and influencers 'sub-contract' their community management to AI bots, and risks for brand partnerships too. Meta does allow its AI characters to be programmed to avoid certain topics, but we've seen so many cases of chatbots going off the rails, making inappropriate suggestions or simply providing inaccurate information that it remains to be seen how useful and reliable they can be on this kind of scale. For now, it seems a reckless strategy for anyone to put their reputation in the hands of an AI avatar whose training it can't fully control. It also makes me feel that the 'social' part of social media is truly coming to an end.
Launching only in the US for now, Meta AI Studio is available at ai.meta.com/ai-studio or via the Instagram app. On Instagram, you simply start a new message and tap 'Create an AI chat' to go to the character creation wizard, choosing your AI character's name, personality and avatar. Meta has an 18-page handbook with some tips.
For more AI tech news, see the new AI tools in Autodesk's Maya and Flame.
