Content creators are busy people. Most spend greater than 20 hours per week create recent content for his or her respective corners of the net. That doesn't leave much time for audience engagement. But Mark Zuckerberg, CEO of Meta, believes AI could solve this problem.
In a interview Together with web star Rowan Cheung, Zuckerberg laid out his vision of a future wherein creatives have their very own bots that capture their personality and “business goals.” Creatives will outsource a number of the community work to those bots to liberate time for other, presumably more vital tasks, Zuckerberg says.
“I believe there's going to be a giant revolution where principally any creative can pull all their information from social media and train these systems to reflect their values and their goals and what they're attempting to do, after which people can interact with it,” Zuckerberg said. “It's going to be almost like this artistic artifact that creatives can create and other people can interact with in alternative ways.”
Zuckerberg's view is widely held in lots of techno-optimistic circles: AI is an innate good since it guarantees to vastly increase the influence of a single person—or organization. (Google has also pitched AI-powered tools for creatives.) But if productivity comes on the expense of the non-public touch, would creatives whose audiences value authenticity really be those to embrace generative AI?
Meta didn't exactly help Zuckerberg's cause, nor did it provide a convincing selling point.
When Meta began rolling out AI-powered bots as a part of its broader Meta AI offensive earlier this yr, it didn't take long for the bots to fall victim to the various pitfalls of today's generative AI technology, most notably hallucinations. The Associated Press observed one bot joining a conversation in a Facebook group for Manhattan mothers and claims it has a baby within the NYC school districtAnother bot offered to offer away a nonexistent camera and air conditioner on a free items exchange forum near Boston.
To be fair, Meta's AI is improving – or so the corporate claims. The latest version, the Llama 3.1 family of models, which can support a spread of features on the tech giant's platforms, is Meta's most mature yet, in response to benchmarks. But hallucinations – and general failures in planning and reasoning – remain an unsolved problem in generative AI, and Meta offers no research breakthroughs here.
It's hard to assume creators putting interaction with their fans within the hands of flawed AI bots. In the interview, Zuckerberg acknowledges that Meta needs to handle “a number of the concerns” about using generative AI and gain users' trust in the long run. This is particularly true since a few of Meta's AI training practices are actively driving creators away from its platforms.
