Stories
Slash Boxes
Comments

SoylentNews is people

SoylentNews is powered by your submissions, so send in your scoop. Only 19 submissions in the queue.
posted by hubie on Saturday March 02, @09:44AM   Printer-friendly
from the if-you-want-to-be-elite-you've-got-to-do-a-righteous-hack dept.

OpenAI has asked a federal judge to dismiss parts of the New York Times' copyright lawsuit against it, arguing that the newspaper "hacked" its chatbot ChatGPT and other artificial-intelligence systems to generate misleading evidence for the case:

OpenAI said in a filing in Manhattan federal court on Monday that the Times caused the technology to reproduce its material through "deceptive prompts that blatantly violate OpenAI's terms of use."

"The allegations in the Times's complaint do not meet its famously rigorous journalistic standards," OpenAI said. "The truth, which will come out in the course of this case, is that the Times paid someone to hack OpenAI's products."

OpenAI did not name the "hired gun" who it said the Times used to manipulate its systems and did not accuse the newspaper of breaking any anti-hacking laws.

[...] Courts have not yet addressed the key question of whether AI training qualifies as fair use under copyright law. So far, judges have dismissed some infringement claims over the output of generative AI systems based on a lack of evidence that AI-created content resembles copyrighted works.

Also at The Guardian, MSN and Forbes.

Previously:


Original Submission

Related Stories

Report: Potential NYT lawsuit could force OpenAI to wipe ChatGPT and start over 20 comments

OpenAI could be fined up to $150,000 for each piece of infringing content:

Weeks after The New York Times updated its terms of service (TOS) to prohibit AI companies from scraping its articles and images to train AI models, it appears that the Times may be preparing to sue OpenAI. The result, experts speculate, could be devastating to OpenAI, including the destruction of ChatGPT's dataset and fines up to $150,000 per infringing piece of content.

NPR spoke to two people "with direct knowledge" who confirmed that the Times' lawyers were mulling whether a lawsuit might be necessary "to protect the intellectual property rights" of the Times' reporting.

Neither OpenAI nor the Times immediately responded to Ars' request to comment.

If the Times were to follow through and sue ChatGPT-maker OpenAI, NPR suggested that the lawsuit could become "the most high-profile" legal battle yet over copyright protection since ChatGPT's explosively popular launch. This speculation comes a month after Sarah Silverman joined other popular authors suing OpenAI over similar concerns, seeking to protect the copyright of their books.

[...] In April, the News Media Alliance published AI principles, seeking to defend publishers' intellectual property by insisting that generative AI "developers and deployers must negotiate with publishers for the right to use" publishers' content for AI training, AI tools surfacing information, and AI tools synthesizing information.

Previously:
Sarah Silverman Sues OpenAI, Meta for Being "Industrial-Strength Plagiarists" - 20230711

Related:
The Internet Archive Reaches An Agreement With Publishers In Digital Book-Lending Case - 20230815


Original Submission

New York Times Sues Microsoft, ChatGPT Maker OpenAI Over Copyright Infringement 51 comments

New York Times Sues Microsoft, ChatGPT Maker OpenAI Over Copyright Infringement

The New York Times on Wednesday filed a lawsuit against Microsoft and OpenAI, the company behind popular AI chatbot ChatGPT, accusing the companies of creating a business model based on "mass copyright infringement," stating their AI systems "exploit and, in many cases, retain large portions of the copyrightable expression contained in those works:"

Microsoft both invests in and supplies OpenAI, providing it with access to the Redmond, Washington, giant's Azure cloud computing technology.

The publisher said in a filing in the U.S. District Court for the Southern District of New York that it seeks to hold Microsoft and OpenAI to account for the "billions of dollars in statutory and actual damages" it believes it is owed for the "unlawful copying and use of The Times's uniquely valuable works."

[...] The Times said in an emailed statement that it "recognizes the power and potential of GenAI for the public and for journalism," but added that journalistic material should be used for commercial gain with permission from the original source.

"These tools were built with and continue to use independent journalism and content that is only available because we and our peers reported, edited, and fact-checked it at high cost and with considerable expertise," the Times said.

AI Threatens to Crush News Organizations. Lawmakers Signal Change Is Ahead 18 comments

Media outlets are calling foul play over AI companies using their content to build chatbots. They may find friends in the Senate:

Logo text More than a decade ago, the normalization of tech companies carrying content created by news organizations without directly paying them — cannibalizing readership and ad revenue — precipitated the decline of the media industry. With the rise of generative artificial intelligence, those same firms threaten to further tilt the balance of power between Big Tech and news.

On Wednesday, lawmakers in the Senate Judiciary Committee referenced their failure to adopt legislation that would've barred the exploitation of content by Big Tech in backing proposals that would require AI companies to strike licensing deals with news organizations.

Richard Blumenthal, Democrat of Connecticut and chair of the committee, joined several other senators in supporting calls for a licensing regime and to establish a framework clarifying that intellectual property laws don't protect AI companies using copyrighted material to build their chatbots.

[...] The fight over the legality of AI firms eating content from news organizations without consent or compensation is split into two camps: Those who believe the practice is protected under the "fair use" doctrine in intellectual property law that allows creators to build upon copyrighted works, and those who argue that it constitutes copyright infringement. Courts are currently wrestling with the issue, but an answer to the question is likely years away. In the meantime, AI companies continue to use copyrighted content as training materials, endangering the financial viability of media in a landscape in which readers can bypass direct sources in favor of search results generated by AI tools.

[...] A lawsuit from The New York Times, filed last month, pulled back the curtain behind negotiations over the price and terms of licensing its content. Before suing, it said that it had been talking for months with OpenAI and Microsoft about a deal, though the talks reached no such truce. In the backdrop of AI companies crawling the internet for high-quality written content, news organizations have been backed into a corner, having to decide whether to accept lowball offers to license their content or expend the time and money to sue in a lawsuit. Some companies, like Axel Springer, took the money.

It's important to note that under intellectual property laws, facts are not protected.

Also at Courthouse News Service and Axios.

Related:


Original Submission

Why the New York Times Might Win its Copyright Lawsuit Against OpenAI 23 comments

https://arstechnica.com/tech-policy/2024/02/why-the-new-york-times-might-win-its-copyright-lawsuit-against-openai/

The day after The New York Times sued OpenAI for copyright infringement, the author and systems architect Daniel Jeffries wrote an essay-length tweet arguing that the Times "has a near zero probability of winning" its lawsuit. As we write this, it has been retweeted 288 times and received 885,000 views.

"Trying to get everyone to license training data is not going to work because that's not what copyright is about," Jeffries wrote. "Copyright law is about preventing people from producing exact copies or near exact copies of content and posting it for commercial gain. Period. Anyone who tells you otherwise is lying or simply does not understand how copyright works."

[...] Courts are supposed to consider four factors in fair use cases, but two of these factors tend to be the most important. One is the nature of the use. A use is more likely to be fair if it is "transformative"—that is, if the new use has a dramatically different purpose and character from the original. Judge Rakoff dinged MP3.com as non-transformative because songs were merely "being retransmitted in another medium."

In contrast, Google argued that a book search engine is highly transformative because it serves a very different function than an individual book. People read books to enjoy and learn from them. But a search engine is more like a card catalog; it helps people find books.

The other key factor is how a use impacts the market for the original work. Here, too, Google had a strong argument since a book search engine helps people find new books to buy.

[...] In 2015, the Second Circuit ruled for Google. An important theme of the court's opinion is that Google's search engine was giving users factual, uncopyrightable information rather than reproducing much creative expression from the books themselves.

[...] Recently, we visited Stability AI's website and requested an image of a "video game Italian plumber" from its image model Stable Diffusion.

[...] Clearly, these models did not just learn abstract facts about plumbers—for example, that they wear overalls and carry wrenches. They learned facts about a specific fictional Italian plumber who wears white gloves, blue overalls with yellow buttons, and a red hat with an "M" on the front.

These are not facts about the world that lie beyond the reach of copyright. Rather, the creative choices that define Mario are likely covered by copyrights held by Nintendo.

AI Story Roundup 27 comments

[We have had several complaints recently (polite ones, not a problem) regarding the number of AI stories that we are printing. I agree, but that reflects the number of submissions that we receive on the subject. So I have compiled a small selection of AI stories into one and you can read them or ignore them as you wish. If you are making a comment please make it clear exactly which story you are referring to unless your comment is generic. The submitters each receive the normal karma for a submission. JR]

Image-scraping Midjourney bans rival AI firm for scraping images

https://arstechnica.com/information-technology/2024/03/in-ironic-twist-midjourney-bans-rival-ai-firm-employees-for-scraping-its-image-data/

On Wednesday, Midjourney banned all employees from image synthesis rival Stability AI from its service indefinitely after it detected "botnet-like" activity suspected to be a Stability employee attempting to scrape prompt and image pairs in bulk. Midjourney advocate Nick St. Pierre tweeted about the announcement, which came via Midjourney's official Discord channel.

[...] Siobhan Ball of The Mary Sue found it ironic that a company like Midjourney, which built its AI image synthesis models using training data scraped off the Internet without seeking permission, would be sensitive about having its own material scraped. "It turns out that generative AI companies don't like it when you steal, sorry, scrape, images from them. Cue the world's smallest violin."

[...] Shortly after the news of the ban emerged, Stability AI CEO Emad Mostaque said that he was looking into it and claimed that whatever happened was not intentional. He also said it would be great if Midjourney reached out to him directly. In a reply on X, Midjourney CEO David Holz wrote, "sent you some information to help with your internal investigation."

[...] When asked about Stability's relationship with Midjourney these days, Mostaque played down the rivalry. "No real overlap, we get on fine though," he told Ars and emphasized a key link in their histories. "I funded Midjourney to get [them] off the ground with a cash grant to cover [Nvidia] A100s for the beta."

Midjourney stories on SoylentNews: https://soylentnews.org/search.pl?tid=&query=Midjourney&sort=2
Stable Diffusion (Stability AI) stories on SoylentNews: https://soylentnews.org/search.pl?tid=&query=Stable+Diffusion&sort=2

This discussion was created by hubie (1068) for logged-in users only, but now has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
  • (Score: 3, Funny) by ElizabethGreene on Sunday March 03, @02:08AM (5 children)

    by ElizabethGreene (6748) Subscriber Badge on Sunday March 03, @02:08AM (#1347160) Journal

    It's possible to goal seek in prompt engineering; If the Times did this they need to share the prompt they used. With that, and the date they did it, OpenAI should be able to load up that model state and run it a few million times to see if it's true.

    • (Score: 3, Touché) by Freeman on Monday March 04, @07:32PM (4 children)

      by Freeman (732) on Monday March 04, @07:32PM (#1347336) Journal

      That way they can patch it and be "in compliance", because "honest we didn't copy the entire internet".

      --
      Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
      • (Score: 3, Funny) by ElizabethGreene on Tuesday March 05, @12:34AM (3 children)

        by ElizabethGreene (6748) Subscriber Badge on Tuesday March 05, @12:34AM (#1347384) Journal

        If they, as you say, copied the entire internet, the power of their compression algorithm would be astonishing.

        My point is that the conditions where they generated the text in the complaint matters. It's possible to goal seek with prompt engineering. You take a prompt and some samples of desired output, and then permute the prompt to get the output you want. This is one of the primary techniques to get an AI that does what you want without spending a pile of money on retraining. I posit there's a big difference between a prompt like "Write a story in the style of the New York Times about the collapse in value of Taxi medallions", running it ten times, and getting their example text vs. a specialized prompt goal seek engineered for a specific of output run a few million times.

        That brings up an interesting theoretical question. Would it be copyright infringement if a million monkeys banging on keyboards for a million years recreated a page of a Disney script? If not, what is the number of monkeys (or AI api calls) that would cause it to be an infringement?

        The conditions matter.

        • (Score: 2, Informative) by Anonymous Coward on Tuesday March 05, @07:30AM

          by Anonymous Coward on Tuesday March 05, @07:30AM (#1347414)

          Sure. And lossy compression can still be copyright infringement.

          Heck in the USA even if it's not compression at all it can still be copyright infringement:
          https://en.wikipedia.org/wiki/Rogers_v._Koons#Opinion_of_the_Court [wikipedia.org]

        • (Score: 2) by Freeman on Tuesday March 05, @02:16PM (1 child)

          by Freeman (732) on Tuesday March 05, @02:16PM (#1347446) Journal

          I sure hope Disney's copyright is up well before 1 million years comes around.

          --
          Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
(1)