Latest / Elon Musk Podcast / NY Times vs. Open AI

Transcript

NY Times vs. Open AI · Elon Musk Podcast

• 0:01
  Hey everybody, welcome back to the Elon Musk Podcast.
• 0:05
  This is a show where we discuss the critical Crossroads, The
• 0:08
  Shape, SpaceX, Tesla X, The Boring Company, and Neuralink,
• 0:13
  and I'm your host, Will Walden. The New York Times has filed A
• 0:17
  lawsuit against Open AI and Microsoft, alleging the
• 0:21
  unauthorized use to millions of its articles to train and
• 0:24
  operate ChatGPT. Now this legal action is the
• 0:27
  most recent among several filed by creators and publishers
• 0:30
  including Sarah Silverman and author George RR Martin amongst
• 0:34
  others. This is against tech companies
• 0:37
  for using their work to develop large language AI models without
• 0:40
  their permission. A central to these lawsuits is
• 0:43
  the practice of scraping, which involves collecting vast amounts
• 0:47
  of Internet data to train AI models like ChatGPT.
• 0:51
  Web crawlers designed to index and download web content are
• 0:55
  increasingly feeding AI models, raising concerns among creative
• 0:59
  content creators about copyright infringement and fair
• 1:02
  compensation. The New York Times claims its
• 1:05
  content was significantly used in the Common Crawl data set,
• 1:09
  which Open AII has admitted to using for training earlier
• 1:13
  versions of Chan GBT. But legal experts are divided on
• 1:17
  whether using Internet data falls under fair use.
• 1:21
  That's the commercial use is a key in consideration.
• 1:25
  Commercial use is a key consideration in determining
• 1:28
  fair use. Now, many AI companies,
• 1:30
  initially nonprofits eventually develop profitable products like
• 1:35
  Open AI websites, have started blocking web crawlers to protect
• 1:39
  their content. Now there's two methods to do
• 1:41
  this. One's based on mutual respect
• 1:43
  and another uses technology to identify and block bad behavior.
• 1:48
  Bots that differ from human users and the reduction in
• 1:51
  accessible data for web crawlers could benefit content creators
• 1:55
  but might also hinder other users like researchers In the
• 2:00
  past, web scrapers were used to collect data about competitors
• 2:04
  and some people use them still for that.
• 2:06
  But also you can get tracking and privacy data from these
• 2:11
  trackers. And now there's an increased
• 2:13
  reliance on web crawling for archiving digital content.
• 2:17
  This modern technique captures online primary sources,
• 2:20
  preserving them as historical records, and major publishers
• 2:24
  have engaged in discussions with Open AI Now about licensing
• 2:27
  content for AI training. However, reaching agreement on
• 2:30
  pricing and terms has been challenging, indicating a
• 2:33
  complex negotiating landscape, and confidential talks have been
• 2:37
  ongoing between top US media companies and Open AI recently.
• 2:40
  Organizations like Ghana News Corp and IAC have been part of
• 2:44
  these discussions, according to sources very familiar with these
• 2:47
  negotiations. Now, Microsoft, who's a huge
• 2:50
  investor in Open AI with millions of dollars invested,
• 2:53
  has also participated in these talks, and the talks have been
• 2:56
  complicated by the rapid development of AI applications,
• 2:59
  raising important questions about the future of the media
• 3:02
  industry. Open AI has expressed respect
• 3:04
  for content creators, rights, and the need for mutually
• 3:07
  beneficial collaborations, as indicated in their deals with
• 3:10
  The Associated Press and Axel Springer.
• 3:13
  The media industry, having previously lost significant
• 3:15
  advertising revenue to tech giants, is cautious about
• 3:18
  undervaluing their content in deals with AI companies.
• 3:21
  There's a concern about AI applications potentially
• 3:24
  spreading misinformation by inaccurately citing articles.
• 3:28
  Some news organizations have successfully negotiated deals
• 3:31
  with Open AI, like The Associated Press and Axel
• 3:33
  Springer. Like I said before, however,
• 3:35
  companies like Bloomberg and the Washington Post have opted to
• 3:38
  focus on their own AI strategies instead of collaborating with
• 3:42
  Open AI Now. Despite these tensions, though
• 3:44
  some industry executives acknowledge the potential
• 3:47
  benefits of AI for journalism, the mutual dependency between
• 3:50
  news organizations and AI firms shows that the need for a
• 3:54
  balance and swift resolution for these disputes is needed.
• 3:58
  The lawsuit underscores the growing tension between the
• 4:01
  media industry and AI tech as well, potentially reshaping the
• 4:04
  news landscape. And Microsoft and Open AI are
• 4:07
  accused of using copyright content to train AI services
• 4:11
  like ChatGPT allegedly causing significant financial damages.
• 4:16
  Microsoft and Open AI have been silent in response to the
• 4:19
  lawsuit. The case represents a major
• 4:21
  challenge to Open AI's practice of scraping web content.
• 4:24
  This is the common practice for ChatGPT since its debut, and the
• 4:28
  company has attempted to secure licensing deals with publishers
• 4:31
  to address all these issues. And now Open AI faces multiple
• 4:35
  lawsuits from various content producers highlighting this
• 4:38
  complex legal terrain. That's surrounding AI and
• 4:41
  copyright right now, and the outcome of these cases could set
• 4:44
  an important precedent for large language models and its
• 4:49
  interaction with content creators.
• 4:51
  And Microsoft is Open AI's largest supporter.
• 4:55
  It's integrated the startups AI tools into its products, and the
• 4:58
  lawsuit alleges that Microsoft's use of the New York Times
• 5:01
  content has significantly boosted its market value.
• 5:05
  Now, the New York Times spokesperson also emphasized the
• 5:07
  legal requirement for obtaining permission before using their
• 5:11
  work for commercial purposes, A requirement they allege
• 5:14
  Microsoft and Open AI have not met.
• 5:17
  And the resolution of this case could have significant
• 5:20
  implications for the future of AI in relation to copyrighted
• 5:24
  content. Hey, thank you so much for
• 5:27
  listening today. I really do appreciate your
• 5:29
  support. If you could take a second and
• 5:30
  hit the subscribe or the follow button on whatever podcast
• 5:33
  platform that you're listening on right now, I greatly
• 5:36
  appreciate it. It helps out the show
• 5:38
  tremendously and you'll never miss an episode, and each
• 5:41
  episode is about 10 minutes or less to get you caught up
• 5:45
  quickly. And please, if you want to
• 5:47
  support the show even more, go to patreon.com/stage Zero and
• 5:53
  please take care of yourselves and each other.
• 5:55
  I'll see you tomorrow.