you can utilize your data to train AI – Zuckerberg

According to a quotation by Zuckerberg, “individual creators or publishers tend to overestimate the value of their specific content.” Stated differently, you can utilize your data to train AI, even if it’s probably worthless.

Fair Play AI companies have purged vast volumes of content at will in order to train their AI models.

This strategy has grown increasingly problematic over time, resulting in numerous lawsuits from copyright holders, including authors, artists, publishers, and major record companies.

In an interview with The Verge, Zuckerberg went so far as to say that “individual creators or publishers tend to overestimate the value of their specific content,” which is likely going to be devastating to the faces of those whose labor is being exploited to build these models.

Zuckerberg claims that if writers decide not to have their work scraped, “we just wouldn’t use their content”.

It’s not like that will make a big difference in how this thing works out,” he said.

His tone deaf comments are indicative of a position that more and more tech leaders in the AI space are taking. Companies like OpenAI have repeatedly argued that the fair use approach, which allows for the restricted use of intellectual content under US law, justifies the scraping of copyrighted data.

“When you put something out in the world, to what degree do you still get to control it and own it and license it?” According to Zuckerberg at The Verge. “I think that all these things are basically going to need to get relitigated and rediscussed in the AI era.”

Hand it on to me. Zuckerberg is by no means the first tech leader to argue that AI systems should be allowed to replicate things that are legally protected. During a 2023 hearing, OpenAI CEO Sam Altman informed lawmakers that content owners ought to be pleased to “get significant upside benefit” from having their work used to train his company’s algorithms.

The CEO of Microsoft AI, Mustafa Suleyman, also made the rather absurd argument earlier this year that any content that is posted “on the open web” qualifies as “freeware” and should be available for use without restriction. This is in violation of US copyright laws, as The Verge notes.

Journalists are upset that AI companies have thrown money at content providers by signing them up for sponsored agreements in an attempt to placate them.

Zuckerberg asserts that content owners just aren’t aware of how little value their work has. Because of this, Meta has a history of prohibiting news organizations from countries where new laws mandating that the company pay for links have been suggested.

The Verge quoted him as saying, “We pay for content when it’s valuable to people.”

We will simply not pay for information that isn’t valuable to users. You will probably notice a similar dynamic with AI, in my opinion. Consequently, news organizations from countries where new legislation requiring the company to pay for links have been proposed have historically been prohibited from using Meta.

We pay for material when it’s valuable to people, he told The Verge. In cases when content isn’t valuable to users, we just won’t pay for them. It is my belief that AI will likely exhibit a comparable dynamic.”