The boring AI that’s saving a moribund Facebook
Facebook has been written off as a network for the elderly and out of touch, but it is learning from its arch-rival TikTok.
When Facebook reported its first-ever decline in active users last year, the Washington Post said the company’s dream was “dead”. CNBC reported its stock was in a “death spiral”. The Atlantic said ”the age of social media is ending”.
And then on Wednesday its boss, Mark Zuckerberg, revealed that users had been spending 7 per cent more time on Facebook this year than the one before. The number of daily users rose 5 per cent to 2.09 billion for the quarter.
These new users might join the app for a host of reasons, but they’re staying for the intrusive videos and seemingly irrelevant posts that now populate users’ Facebook feeds. In other words: the TikTokification of Facebook is working. And that says something about which forms of artificial intelligence are real, and which bits are driven by hype.
The news: Meta Platforms, which also owns WhatsApp and Instagram, said its quarterly revenue rose 23 per cent to $US34.15 billion ($53.6 billion) from a year earlier to a mixed market reaction. Net income more than doubled to $US11 billion or $US4.39 a share, cent over last year, when its workforce was bloated. In a boon to other media businesses, Meta showed ads were holding up despite interest rate rises, though it flagged softer spending for the ongoing quarter “correlating” with the beginning of the Hamas-Israel conflict. And after letting TikTok steal a march on it with its feed of endless algorithmically selected videos, Meta is back in the game.
What’s driving it: Facebook has had Reels, the short-form, vertical videos that originally debuted in Instagram and were inspired by TikTok, for two years. But the videos and posts Facebook shows users are getting better at making users stay. Where Facebook once only showed users content from people and pages they followed, now it deduces user’s preferences via an AI system that figures out if they want cooking content, gym tips, alt-right comedy or something else entirely. That kind of behind-the-scenes AI is much more useful to businesses right now than the showier generative tools.
The quote: Zuckerberg said he was confident generative AI – the kind that lets users create images or passages of text from a brief prompt – was definitely “a thing”. But he was more specific about the benefits of the unseen AI that determines what users see, which had boosted time spent on Facebook by 7 per cent and 6 per cent on Instagram. “This technology has less hype right now than generative AI, but it’s also improving very quickly,” Zuckerberg said.
Our take: Meta has to be careful with how it uses artificial intelligence in its algorithms. Just yesterday it was hit with a bipartisan lawsuit from 41 US states claiming that its products are addictive and harmful for children. “Meta has harnessed powerful and unprecedented technologies to entice, engage, and ultimately ensnare youth and teens,” the main lawsuit claims. TikTok has been a boon for Meta’s lobbying and public image, allowing it to present itself as a comparatively responsible corporate citizen. As it apes TikTok’s features, that distinction could start to look like splitting hairs.
And another thing: Meta’s Reality Labs division, which makes its virtual and augmented reality goggles, lost $US3.7 billion last quarter. Zuckerberg’s original pitch for the division was that it would build the Metaverse, something he was so enthused about that he renamed his company. But his stance has changed to claiming that “the ability to deliver AI through smart glasses may end up being a killer use case”. When a business pivots between two hyped technologies that rapidly, it pays to be sceptical.
The Breakdown is a regular column that unpacks and explains big stories from the tech world.
Introducing your Newsfeed
Follow the topics, people and companies that matter to you.
Find out moreRead More
Latest In Technology
Fetching latest articles