Content creators are busy people. Most spend 20+ hours a week creating new content for their corner of the web. That doesn’t leave much time for audience engagement. But Mark Zuckerberg, CEO of Meta, believes AI can solve this problem.
In an interview with internet personality Rowan Chan, Zuckerberg laid out his vision of a future in which creators have their own bots that capture their personality and “business goals.” He said creators could hand off some of their community outreach to these bots, freeing up their time for other, perhaps more important tasks.
“I think we’re going to see a big shift where basically all creators are taking all of their information from social media and training these systems to reflect their values and their purpose and what they’re trying to do, and then people are going to be able to interact with that,” Zuckerberg said. “This is like a piece of art that creators create, and people are going to be able to interact with it in different ways.”
Zuckerberg’s perspective is common among many tech optimists: AI is an inherently good thing that promises to vastly expand the impact any one person or organization can have. (Google is also touting AI-powered tools for creators.) But will creators whose audiences value authenticity really embrace generative AI if productivity comes at the expense of the personal touch?
It doesn’t help Zuckerberg’s position, but Meta hasn’t made a particularly strong pitch.
When Meta began rolling out AI-powered bots as part of its broader Meta AI push earlier this year, it didn’t take long for the bots to fall prey to many of the pitfalls of today’s generative AI technology, particularly hallucinations. The Associated Press saw one bot barge into a conversation in a Facebook group for Manhattan mothers, claiming to have children in the New York City school district. Another bot offered to give away nonexistent cameras and air conditioners on a free-item swap forum near Boston.
In fairness, Meta’s AI is improving — at least that’s what the company claims. Its latest release, the Llama 3.1 model family, which will power a range of features on the tech giant’s platform, is the most sophisticated yet, as measured by benchmarks. But hallucinations and common mistakes in planning and reasoning remain open problems in generative AI, and Meta offers no research breakthroughs there.
It’s hard to imagine creators putting their trust in a flawed AI bot to interact with their fans. In the interview, Zuckerberg acknowledged that Meta needs to “alleviate some of the concerns” about using generative AI and earn users’ trust over the long term. This is especially true given that some of Meta’s AI training methods actively drive creators away from the platform.
