Getty Images/iStockphoto

Meta introduces new AI assistant and studio

The social media giant introduces new tools that incorporate generative AI capabilities in Facebook, Instagram and WhatsApp. Some features could be helpful for small businesses.

Facebook's parent company, Meta, on Wednesday introduced a series of what it calls experiences across different apps and devices aimed at both consumers and some businesses.

Among the new features are AI stickers, a way for consumers to generate new stickers for chats and stories. AI stickers uses technology from Meta's open source Llama 2 large language model and Meta's new foundation model for image generation, Emu, an AI tool that turns text prompts into stickers.

Stickers will be available in WhatsApp, Messenger, Instagram and Facebook stories over the next month.

Another new tool is Meta AI, an assistant for users of WhatsApp, Messenger, Instagram and the upcoming Meta-powered Ray-Ban smart glasses and the Quest 3 mixed reality headset. Available now, Meta AI is powered by a custom model based on Llama 2. It also uses Meta's search partnership with Bing and includes an image generation tool.

For coders and noncoders, Meta said it will roll out an AI studio for users outside the company's platforms. Using the studio, which is now in alpha, businesses will be able to create their own AI tools for branding and customer service.

The enterprise angle

While much of what Meta introduced is consumer-facing, there is also an enterprise pitch, said Arun Chandrasekaran, an analyst at Gartner.

There is clearly the small business and enterprise side of this as well, primarily in terms of productivity, better communication and user engagement.
Arun ChandrasekaranAnalyst, Gartner

"There is clearly the small business and enterprise side of this as well, primarily in terms of productivity, better communication and user engagement," Chandrasekaran said.

If Meta were to include some generative AI features within apps like WhatsApp, such as summarization of texts or even translation, that could be helpful not only for consumers but also to some small businesses, he added. Those features were not part of Wednesday's product introduction at Meta's Connect 2023 virtual conference.

Another enterprise application could be around the Emu model image generation model, said Forrester analyst Rowan Curran.

"Depending on how that model is trained and the data sets that are used for it, it may be another option for folks beyond the existing providers and open source models out there," Curran said.

Also, AI studio could also have enterprise-reaching capabilities if the chatbot functionality becomes more extensible beyond the Facebook platform.

Moreover, Meta's use of AI models, such as Llama 2 and Segment Anything, shows that while some of its offerings today may not have an immediate enterprise impact, they are still notable in terms of their influence on the tech world beyond the Meta social media platform universe.

"The models that they are continuing to develop and release are certainly going to have significant impacts on the broader market, not just within Meta and for their releases and products," Curran said.

Keeping users engaged

However, Meta's goal is ultimately to keep users on its social media properties, using AI to keep audiences engaged and even help small businesses think about their social media strategy, Chandrasekaran said.

Meta is not the first social media company to aim AI technology at boosting user productivity, engagement and experience. For example, Snapchat introduced its Snap AI assistant earlier this year as part of a partnership with OpenAI.

It's expected that Meta's competitors, such as X, will follow the same example, Chandrasekaran said.

Also, at Connect, Meta revealed that it has partnered with a number of influencers and celebrities and created AI icons and characters from them.

Esther Ajao is a TechTarget Editorial news writer covering artificial intelligence software and systems.

Next Steps

Meta's new silicon shows a growing trend for AI hyperscalers

Dig Deeper on AI technologies