Latest / Elon Musk Podcast / OpenAI Unleashes Groundbreaking AI Model Amidst Controversy and Legal Battles!

Transcript

OpenAI Unleashes Groundbreaking AI Model Amidst Controversy and Legal Battles! · Elon Musk Podcast

• 0:01
  Hey everybody, welcome back to the Elon Musk.
• 0:04
  Podcast this. Is a show where we discuss.
• 0:07
  The Critical. Crossroads that shape SpaceX,
• 0:09
  Tesla. X.
• 0:11
  The Boring Company and Neuralink.
• 0:13
  I'm your host, Will Walden. Open AI announced on Tuesday
• 0:18
  that it has commenced training its newest AI model to replace
• 0:22
  the current GPT for tech, which powers the powerful and popular
• 0:26
  Chant GPT Chat bot. This new development is part of
• 0:30
  Open a is broader goal to achieve artificial intelligence.
• 0:34
  It's artificial general intelligence, actually, a Gia
• 0:38
  concept that envisions machines with cognitive capabilities akin
• 0:42
  to human brains. In addition to this major
• 0:44
  update, Open AI revealed the formation of a new committee
• 0:47
  dedicated to evaluating the safety and security of its AI
• 0:51
  products. This committee, which includes
• 0:53
  CEO Sam Altman, will conduct a 90 day review of the company's
• 0:57
  technology and release the recommendations thereafter, he
• 1:00
  said. We welcome a robust debate at
• 1:03
  this important moment, emphasizing their openness to
• 1:06
  discuss anything regarding safety measures and ethical
• 1:10
  considerations of their AI developments.
• 1:13
  Now based in San Francisco, Open AI is a leading player in the
• 1:16
  artificial intelligence sector known for widely used tools such
• 1:20
  as ChatGPT and also Doll E. Now these technologies has seen
• 1:25
  significant adoption with large corporations like Microsoft, a
• 1:29
  major financial backer of Open AI, integrating them into their
• 1:32
  operations. Now, the announcements come at a
• 1:35
  turbulent time for Open AI as the company faces a series of
• 1:38
  challenges, including lawsuits, high profile staff departures,
• 1:42
  and concerns over the potential misuse of AI tech.
• 1:45
  And among these issues are fears that AI could facilitate
• 1:49
  election disinformation, job displacement, and copyright
• 1:53
  infringement. Recently, Open AI introduced an
• 1:56
  update to its AI technology GPT 4 O or GPT 4 Omni enhancing Chat
• 2:03
  GP TS capabilities to respond by voice.
• 2:06
  Now, this update has not been without controversy, as actress
• 2:10
  Scarlett Johansson accused Open AI of creating a voice really
• 2:13
  similar to hers without permission.
• 2:15
  We have an episode all about that a few episodes ago, so
• 2:18
  please check that out after you listen to this one.
• 2:21
  Now, Open AI responded to Johansson's accusations by
• 2:23
  clarifying that the voice was not hers and that a voice actor
• 2:26
  had been hired and paid substantially for that.
• 2:30
  Despite this, Johansson has threatened legal action and this
• 2:33
  is an ongoing concern about the ethical use of AI generated
• 2:37
  content in the future. Now, the company is already
• 2:40
  facing multiple lawsuits from news organizations, including
• 2:42
  The New York Times, alleging that Open AIS methods of
• 2:46
  training its chatbot technology infringe on copyrights.
• 2:50
  These lawsuits challenge the practice of using vast amounts
• 2:53
  of online data, such as news articles and other digital
• 2:57
  content, to train AI models. And the departure of Open AI Co
• 3:00
  founder Ilya Sutskever in May has added to the turmoil
• 3:06
  internally. Now they led the team focused on
• 3:10
  the societal threats of AI, resigned amid criticism that
• 3:14
  Open AI was prioritizing profit over addressing the dangers
• 3:18
  posed by the technology. And the resignation followed a
• 3:21
  contentious episode in November when he and other board members
• 3:25
  voted to oust Co founder Sam Altman.
• 3:27
  And Altman was reinstated shortly thereafter, leading to
• 3:31
  the resignation of Scutskever and other dissenting directors
• 3:36
  from the board. Now, the newly formed safety
• 3:38
  committee chaired by Brett Taylor, Cora CEO Adam D'Angelo,
• 3:42
  former Sony and her team executive Nicole Seligman and
• 3:46
  Altman will focus on refining open AI safety and security
• 3:49
  protocols. This move comes shortly after
• 3:52
  former board members called for increased regulation on AI
• 3:55
  technologies. Now, these regulatory calls were
• 3:58
  echoed in a recent article in The Economist, where former Open
• 4:02
  AI board members argue that AI companies like Open AI cannot be
• 4:06
  trusted to self regulate effectively.
• 4:08
  This has intensified the debate over how to best manage the
• 4:11
  rapid advancements in AI technology, and Open A IS
• 4:14
  aggressive push towards developing more advanced AI
• 4:16
  models aims to outpace its competitors while simultaneously
• 4:20
  addressing the criticisms surrounding a IS potential
• 4:23
  dangers. The company's ambition to
• 4:25
  achieve AGI remains a contentious topic within the
• 4:28
  tech industry and beyond, and the latest iteration of GPT 4,
• 4:32
  released in March of 2023, has significant enhancements.
• 4:37
  Indeed. Chat bots and other software
• 4:40
  applications performing complex tasks, of course, answering
• 4:44
  questions, generating text, analyzing data.
• 4:47
  And the newest updated version, GTP four O, adds voice
• 4:52
  interaction capabilities, further expanding the potential
• 4:54
  uses of the technology. Now, if you're on a platform
• 4:57
  that can comment on this podcast, please let me know if
• 5:00
  you've used the latest version of GPT 4 O for Omni.
• 5:05
  Actually everybody calls it four O though, so I've been using it
• 5:09
  recently to do some some coding work and apparently it's like
• 5:15
  it's it's pretty similar to regular GPT 4, but I can use
• 5:18
  voice commands to allow it to configure code for me instead of
• 5:23
  typing everything out. It can do a lot of the work with
• 5:27
  just me talking to it, which is pretty incredible actually.
• 5:30
  Now trending these AI models like GPT 4 O involves extensive
• 5:34
  data analysis including sounds, images, videos, and textual
• 5:39
  content from various sources. This process can take gears,
• 5:43
  possibly with additional time required for testing and fine
• 5:46
  tuning before public release. So they've been working on this
• 5:49
  for a very long time and we just got a peek at it.
• 5:52
  And this isn't even the final version of this.
• 5:54
  They're going to keep iterating on it and making it better.
• 5:57
  They're trying to blow away the competition, especially Grok and
• 6:01
  Google's GPT models as well. They're going to it.
• 6:06
  It's a stiff competition out there right now.
• 6:08
  So everything they include in an update is going to be pretty
• 6:13
  ground breaking going forward. Multimodal actually is what
• 6:17
  they're what they're working on right now.
• 6:19
  They're they're very multimodal at this time, but they're
• 6:23
  pushing for more. So possibly in the future you
• 6:26
  can upload videos and have them have it more fit, have it do any
• 6:31
  sort of any sort of work on it that you need done.
• 6:35
  I'm not sure what they're going to be doing, but hopefully they
• 6:36
  can do some video editing with the tool.
• 6:39
  I think that would be really great, like an AI generated
• 6:42
  video editing tool would be very interesting for open AI.
• 6:46
  But this newest model, it's expected to be completed in the
• 6:50
  next nine months to a year and it'll undergo rigorous safety
• 6:53
  assessments before we get our hands on it by the committee.
• 6:56
  And this committee will ensure that the model meets stringent
• 6:59
  safety and security standards before it's made widely
• 7:03
  available. Now, the leadership of John
• 7:05
  Shulman, a Co founder and former head of the team, the created
• 7:09
  ChatGPT, will be crucial in guiding Open A is long term
• 7:12
  safety research. This new committee will oversee
• 7:16
  Shulman's work and help establish robust policies to
• 7:19
  mitigate technological risks. And it is open AI continues down
• 7:25
  this path of complex challenges. You're going to keep pushing
• 7:28
  boundaries because they're a private company and they need to
• 7:32
  make money. And the success of its new model
• 7:35
  and the effectiveness of its safety measures, It's going to
• 7:39
  be the focus over the next 6 to 9 months, and we'll keep you
• 7:42
  updated here on the show. Hey, thank you so much for
• 7:52
  listening today. I really do appreciate your
• 7:54
  support. If you could take a second and
• 7:56
  hit the subscribe or the follow button on whatever podcast
• 7:59
  platform that you're listening on right now, I greatly
• 8:02
  appreciate it. It helps out the show
• 8:03
  tremendously and you'll never miss an episode.
• 8:06
  And each episode is about 10 minutes or less to get you
• 8:10
  caught up quickly. And please, if you want to
• 8:12
  support the show even more, go to patreon.com/stage Zero.
• 8:18
  And please take care of yourselves and each other, and
• 8:21
  I'll see you tomorrow.