Didn't get a chance to hear yesterday's oral arguments in the @nytimes AI copyright lawsuit against @OpenAI & @Microsoft
? Here's a recap by @SPGuaglione and me, which has some key points both sides made in federal court. (Link to the @Digiday story in the thread.)
@martyswant
-
NYT AI Copyright Lawsuit Against OpenAI Microsoft Oral Arguments
By
–
-
NYT v. OpenAI Oral Arguments: Legal Implications Analysis
By
–
Anyone hear today’s oral arguments in NYT v. OpenAI and want to share a quick POV before 3pm ET? Curious what everyone thought of both sides’ points, the case law cites, judge’s questions, potential outlook/implications, etc.
-
Copyright Law and AI Training Data: Copying as Injury
By
–
“The injury is the copying.” said a lawyer for the Center for Investigative Reporting, adding that U.S. copyright law only requires someone to prove copying rather than harm.
-
AI Copyright Infringement Liability Without Traceability Protection
By
–
Judge: What’s the injury? Lieberman: “You’re leaving people open for massive copyright infringement without the ability to trace it…It’s like it causes the alarm system in your house to go down.”
-
Copyright Management and Attribution Retraction in AI Outputs
By
–
Lieberman: When copyright management information (CMI) is retracted, outputs will be "either the verbatim language or a summary without the attribution of the new york times or the daily news.”
-
ChatGPT Copyright Trial: Sources Not Cited to Users
By
–
The judge asked plaintiffs to show examples of harm. Steven Lieberman, a lawyer repping NY Daily News and NYT, gave examples of how ChatGPT won't cite the original source when user asks an LLM for info about an old Super bowl or a past presidential debate.
-
AI Search Citations Risk Siphoning Traffic From Original Sources
By
–
Crosby pointed out traditional search engines give blue links to original sources, but AI search provides answers w/ citations to original sources. He worries AI citations could further siphon traffic from original sources by boosting user confidence in answers via citations.
-
NYT Lawyer Distinguishes LLM Outputs from Search Engines
By
–
NYT's lawyer Ian Crosby said LLM outputs are just "the expression of the words and facts, not the underlying facts in the words and the models.” “They’re not next-generation search engines, but answer engines," Crosby said, adding that they're not substitutional.
-
OpenAI Lawyer Clarifies GPT Training Data in NYT Copyright Case
By
–
Judge just asked OpenAI's lawyer if GPT models have NYT content in its database. Lawyer said not doesn't, but then clarified it doesn’t have a database but instead relies on weights. I wonder how NYT's lawyers will respond to that.
-
OpenAI Lawyer Explains LLM Use Case in NYT Trial
By
–
One of OpenAI's lawyers in this NYT v OpenAI trial is explaining LLMs to the judge & said someone can use ChatGPT to make a letter to a landlord "sound less angry." haha I've literally did that when our heat went out last year! (Full circle with out heat broken again today.)