Digital Copyright and AIs

AIs may train on everything. You may not.

May 14, 2024 โ€” In America, AIs have more freedom to learn than humans. This worries me.

Do you want learn at the same library as ChatGPT, Gemini, Grok, or Llama?

Then you must become a criminal.

You have no legal option[1].

*

This sounds ridiculous.

Human made laws that work against humans?

But it is true.

*

The problem is Copyright Law. Specifically, Digital Copyright.

Big corporations have the resources to legally build massive internal libraries of all the world's published information.

Their AIs can legally train on this.

Humans in America have to choose: use a criminal library, or fall behind AIs?

*

A Proposed Fix

Some smart people want to abolish copyright on digital publications.

Among other benefits, this would ensure all Americans have the same freedom to learn as AIs.

*

Does Digital Copyright matter?

Of course, we may not need to update the law.

*

But you never know.

Now might be a good time to set into law equal educational rights between humans and AIs.

While the laws are still set by humans.

โ‚

Notes

[1] Well, there is no legal way for YOU to do it. A billionaire could theoretically afford to buy the "rights" and build their own library as good as a big corporation. But no billionaire is reading this post. Nevertheless, to be precise, only ~756 out of ~333,333,000 Americans ~have a legal way to train on the same materials as AIs.