Meta has created a new team dedicated to building a suite of artificial intelligence-powered tools that could include “AI personas” that are designed to help people. It comes as Twitter owner Elon Musk is rumoured to be building a new research lab to create an “anti-woke” alternative to ChatGPT.
Writing in a Facebook post, Meta CEO Mark Zuckerberg said the new top-level product group would be focused on generative AI to “turbocharge our work in this area”. It makes his company the latest in a long line of Big Tech businesses around the world to commercialise generative AI.
Meta has invested heavily in AI as a by-product of its metaverse push, including publishing a number of open-source models, products and tools widely used by the AI industry. This includes the large language model LLaMA-65B released last week and available open-source for industry labs and researchers. It reportedly performs as well as models such as GPT-3 with fewer parameters.
It seems AI is now becoming a core focus for the company as Zuckerberg works to focus on revenue-generating products. With the rise of AI in the public sphere, including OpenAI’s ChatGPT becoming the fastest-growing consumer product in history, hitting 100 million active monthly users in January, and Bing launching its own AI chatbot, the technology is entering a commercial phase.
Google is also working on its own AI chatbot called Bard and plans to incorporate the technology throughout its product range and Microsoft is adding AI to Teams, Microsoft 365 and Azure. Chinese tech giants, including Tencent, Baidu and Alibaba, are also working on chat products.
Commercialising artificial intelligence
Zuckerberg said Meta has formed the new division “by pulling together a lot of the teams working on generative AI across the company into one group focused on building delightful experiences around this technology into all of our different products”.
The company plans to focus on adding creative and expressive AI tools to its existing products in the short term which will likely see AI filters added to Instagram and generative text in WhatsApp and Facebook. There is likely to be a range of text, image and video tools across the range.
“Over the longer term, we’ll focus on developing AI personas that can help people in a variety of ways,” Zuckerberg wrote. “We’re exploring experiences with text, with images, and with video and multi-modal experiences. We have a lot of foundational work to do before getting to the really futuristic experiences, but I’m excited about all of the new things we’ll build along the way.”
The product team will be led by Ahmad Al Dahle, the current VP of AI/ML and core tech at the social media giant and he will be reporting to Meta’s chief product officer Chris Cox, which the company says will allow it to more rapidly deploy AI in its products than has been the case in the past.
While Microsoft and Google have been bullish about incorporating AI widely into products, Meta has been cautious with the technology previously, focusing on AI as a research play rather than a revenue-generating product. This appears to be changing, although Meta says it will continue to have researchers working on longer-term AI projects outside of the product team, focusing on academic research Meta will publish and share widely.
Musk wants ‘anti-woke’ AI chatbot
Meanwhile, Musk is reported to have been approaching leading AI researchers to discuss creating a new lab designed to build an alternative to ChatGPT, including training new large language models.
His goal is apparently to create a chatbot without the perceived bias shown in text generated by the OpenAI tool. Musk has been critical of OpenAI for putting safeguards and boundaries in place to prevent ChatGPT from producing offensive content and accused them of training it to be woke. He was a co-founder of OpenAI when it was created as a pure not-for-profit research lab, but cut ties with the group as it changed direction to focus on profit.
He is apparently in talks with Igor Babuschkin, an AI expert who recently left DeepMind, the AI company owned by Google-parent Alphabet. DeepMind has its own chatbot built on a large language model called Sparrow that hasn’t been released to the public.
Musk has praised and been critical of ChatGPT describing it as “scary good” in tweets and saying “we are not far from dangerously strong AI” which is also known as AGI or Artificial General Intelligence.
He has previously described AI as “one of the biggest risks to the future of civilisation” but has also said he would rather be alive now to witness AGI than in the past and not be able to see it come about. He said it needs to be regulated and according to Babuschkin, in an interview with The Information, building a chat tool with fewer safeguards isn’t an objective mentioned by Musk, rather the focus is on combating bias.