AI and the Law on FIRE's "So To Speak" Podcast (with Nico Perrino, David Greene [EFF], Alison Schary, and Me)

The title is "Artificial intelligence: Is It Protected by the First Amendment?," but we covered more than that. I much enjoyed it, and I hope you will, too! Here's FIRE's summary:

What does the rise of artificial intelligence mean for the future of free speech and the First Amendment? Who is liable for what AI produces? Can you own a copyright for works produced by AI? Does AI itself violate intellectual property rights when it uses others' information to generate content? What about that Morgan Freeman "deep fake"? And is ChatGPT going to make all of our jobs irrelevant?

Guests:

• Eugene Volokh, professor at UCLA School of Law
• David Greene, senior staff attorney and civil liberties director at the Electronic Frontier Foundation
• Alison Schary, partner at Davis Wright Tremaine

www.sotospeakpodcast.com YouTube: https://www.youtube.com/@freespeechtalk Twitter: https://www.twitter.com/freespeechtalk Facebook: https://www.facebook.com/sotospeakpodcast Instagram: https://www.instagram.com/freespeechtalk/ Email us: [e-mail]

Comments

[Archibald Tuttle]

My first reaction was, "silly question." But after watching EV's opening statement, I changed my mind. It's a good thing to think about the issues now because it reduces to the rights of persons, not the AI.

[Drannab]

AI is a tool. I can use it for art, music, or writing. I’m still the author. While not a lawyer, it’s a tool for those who want to represent themselves. That being said, as a consumer of that tool, I want it to have some sort of certification and insurance that it wouldn’t spout gibberish in a court case.

[Doug Huffman]

AI is a tool. One must be able to detect gibberish to use the tool. One must be able to formulate a good question to avoid gibberish.

ChatGPT is quite capable of error and of intentional untruth.

[Archibald Tuttle]

The law applies to persons, therefore law-given rights apply only to persons.

Does anyone remember the episode of Hot In Clevland where the parrot says, F-U? Would anybody ask about the parrot's 1A rights?

Persons, not persons, that binary choice applies to the law. There is no halfway case.

[AmosArch]

Who is liable for what AI produces? Can you own a copyright for works produced by AI? >>>>>>>>>>>>>>>>>

‘AI generators’ are pretty much a fancy photoshop denoiser. Adobe doesn’t own all works produced on its software so why should it be different with ‘AI’?

>>>>>>>>>>>>>>>>>>> Does AI itself violate intellectual property rights when it uses others’ information to generate content? What about that Morgan Freeman “deep fake”? >>>>>>>>>>>>>>>

Unlike what some people think. ‘AI’ doesn’t directly store and mash together works. Rather it stores statistical stylistic information through the optimization of parameters. Which can be roughly analogized to a person learning a style. As far as I know if you are a human the estate of Andy Warhol can’t pop up and sue you for a million dollars just because you learned how to draw faces from one of his paintings. So for the above, not unless you think a computer arguably doing something thats legal for a human to do should be illegal.

This argument is basically a smokescreen anyway. The antigenerator crowd isn’t afraid of the software scraping. They’re afraid of being replaced. Once the generator companies hire people to provide training content I doubt the controversy is going away.

[Hank Ferrous]

ChatGPT won't make administrative jobs irrelevant, but AI will in time. As much as the white collar crowd dislikes the concept, Prof V, even attorneys will be replaced eventually.