The Dirty Little Secret of AI That More People Should Be Talking About

There’s no question that generative AI has completely transformed the world of content creation. In just a few seconds, a large language model (LLM) such as ChatGPT can create just about any type of content you could possibly ask for, in whatever style you want.

But here’s the thing: AI is only able to do this because it has been trained on human-created content that exists all over the internet. AI essentially takes what has already been created by humans, and then transforms it, according to your prompt. The whole process is so quick and so powerful that it almost seems like magic.

The dirty little secret of AI

But now content creators and web site operators are fighting back. As they see it, the blood, sweat, and tears that they have put into creating content or developing an online business is now being compromised by AI. Very powerful AI bots are training on this content, without even so much as a hat tip or a “thank you.” Isn’t this really just a very clever and high-tech form of plagiarism?

That’s the question that U.S. federal judges are attempting to answer. A group of three authors recently sued Anthropic, the AI company where Amazon is a significant investor. As they see it, Anthropic violated their copyrights on published material, and they deserve some form of compensation. They claim that Anthropic created a multi-billion-dollar business, all by “stealing” thousands of copyrighted works. Even worse, Anthropic hadn’t even paid for a single copy of the books - they simply found a pirated version, and used it for free.

However, that’s just my perspective as a content creator. The U.S. district judge presiding over the case didn’t see it that way. As the judge saw it, the AI bot from Anthropic was simply engaging in “fair use.” Moreover, the judge viewed the creativity and innovation used by the Anthropic bots as “exceedingly transformative.” The LLM did not violate any copyrights, because it didn’t directly reproduce the works for sale.

So, even though the LLM used the books for profit, they didn’t directly plagiarize the work, so everything’s OK? The analogy used in the case is either going to make you laugh or cry. The judge said that the case of the LLM using the books to create new content was similar to “a reader aspiring to be a writer.” If you check out a book from a public library, so that you can study how to become a better writer or author, you don’t pay any money, right? So why should an LLM?

A tightrope situation

As you can tell, this precedent-making case is walking a legal tightrope. From a purely technical perspective, no plagiarism or violation of copyright took place. However, consider the case from the perspective of the authors. Doesn’t it seem obvious that AI is benefiting financially from their work?

Making matters even worse, Anthropic doesn’t seem to be really sympathetic to these concerns. The company says that the purpose of copyright is to enable creativity and to foster technological progress, and isn’t that what its AI is doing? In other words, we should be so grateful that AI is training on the hard work of humans and using it to crank out such creative answers.

Even though this case is turning out to be a big win for AI companies, the public view of AI is becoming more and more negative by the day. Earlier this summer, Anthropic even warned that half of all entry-level white-collar jobs would soon be eliminated, thanks to AI. So, soon, it won’t just be authors and content creators complaining about AI. It will be every single young person who’s gainfully employed right now.