Meta Launching AI-Powered Friends to Talk to on Facebook

When was the last time you logged into Facebook?

Facebook's parent company Meta is about to launch AI-powered chatbots that are meant to lure users back to the mammoth social media platform, the Financial Times reports. Further down the road, the company is apparently working on AI agents in its namesake Metaverse as well.

In other words, your estranged aunt and lonely grandpa are about to be joined by virtual friends to fill the void you left behind when you quit Facebook and never looked back.

According to the report, these chatbots are supposed to be able to chat with the platform's billions of users in a life-like way, taking on various "personas." For instance, an Abraham Lincoln AI could teach you about the abolition of slavery, while a travel advisor could tell you about the latest and greatest surfing destinations.

But whether that will be enough of a draw for bring people who left Facebook years ago remains to be seen. Young people have long dismissed the almost two-decades-old platform as an irrelevant place for "old people." Younger generations have long moved on to other platforms like Meta's Instagram and TikTok.

In short, Facebook has been fighting to maintain a sense of relevancy. Yet active users aren't exactly the most pressing of issues: the company announced last week that it had surpassed a whopping 2.064 billion daily active users, a modest increase over last quarter.

Looking at the bigger picture, introducing AI chatbots is likely an attempt to catch up with Meta's competitors in the AI space. Earlier this year, Meta CEO Mark Zuckerberg told Axios that "incredible breakthroughs" in generative AI "gives us the opportunity to now go take that technology, push it forward, and build it into every single one of our products."

And the company isn't starting from square one. Last month, Meta announced its latest large language model, called Llama 2, that the company could use to power these chatbots.

The chatbots aren't just supposed to boost engagement. They may also hone in even further on users' interests, experts told the FT. In other words, Meta is looking for an AI-enhanced way to better target more relevant content — and, of course, ads.

Those kinds of interactions could have implications on user privacy as well. Facebook already has a poor track record when it comes to user privacy — and throwing loose-lipped and misinformation-spewing chatbots into the mix certainly comes with plenty of inherent risks.

"Once users interact with a chatbot, it really exposes much more of their data to the company, so that the company can do anything they want with that data," Ravit Dotan, an AI ethics adviser at the University of Pittsburgh, told the newspaper.

Then there's the issue of AI chatbots "hallucinating" facts and spreading misinformation. Given Meta's earlier attempts — we're looking at you, BlenderBot 2 — there's a chance, but no guarantee, that the company will play it safe and build out ways to screen user input.

However, there's a lot we still don't know about Facebook's foray into AI chatbots. One insider told the FT that Zuckerberg is "spending all his energy and time on ideating about" avatar chatbots that could chat with users inside the metaverse in the long term.

Will the company play it safe and throw its weight behind finding new ways to keep these chatbots in check? Will the move even entice users to come crawling back to a platform they left long ago?

We've also already seen the spectacular levels of hype surrounding AI chatbots wane as of late, meaning that Facebook could be late to the party. Even Facebook's own AI guru and machine learning pioneer Yann Lecun said earlier this summer that large language models are a passing fad.

All told? We'll have to wait and see. But we'll be watching.

More on Facebook: Facebook's Chief AI Scientist Says LLMs Are Just a Passing Fad