Latest / Bloomberg Intelligence / OpenAI Says It’s Working With Anthropic, Google on AI Safety

Transcript

OpenAI Says It’s Working With Anthropic, Google on AI Safety · Bloomberg Intelligence

• 0:02
  Speaker 1: Bloomberg Audio Studios.
• 0:04
  Speaker 2: Podcasts.
• 0:05
  Speaker 3: Radio. News.
• 0:09
  Speaker 1: You're listening to the Bloomberg Intelligence Podcast. Catch us live
• 0:13
  Speaker 1: weekdays at 10 a.m. Eastern on Apple CarPlay and Android
• 0:16
  Speaker 1: Auto with the Bloomberg Business App. Listen on demand wherever
• 0:19
  Speaker 1: you get your podcasts or watch us live on YouTube.
• 0:24
  Speaker 2: Open AI, the story now is getting political, it seems like, Scarlett.
• 0:28
  Speaker 2: I mean, we have President Trump out saying... to these
• 0:30
  Speaker 2: AI companies, continue to spend, continue to invest.
• 0:33
  Speaker 4: There's nothing to worry about. Move on, move on, move on.
• 0:35
  Speaker 2: Nothing to worry about. And I think he raises a
• 0:37
  Speaker 2: very good point, which is this is kind of competitive race,
• 0:40
  Speaker 2: China versus the West, China versus the U.S. So to
• 0:44
  Speaker 2: the extent you do maybe pull back for whatever reason,
• 0:47
  Speaker 2: you do run the risk of maybe falling behind.
• 0:49
  Speaker 4: But it's a sensitive time to be saying all this
• 0:51
  Speaker 4: because lawmakers are gearing up for the midterm elections and
• 0:55
  Speaker 4: they need to be at least listening to their constituents.
• 0:58
  Speaker 4: And a lot of constituents have a lot of concerns
• 1:00
  Speaker 4: about AI and what it's going to do to the economy,
• 1:03
  Speaker 4: to their jobs, to their livelihood.
• 1:04
  Speaker 2: All right, let's see what investors are saying. Matthew Bloxham
• 1:06
  Speaker 2: joins us here. He's a Bloomberg Intelligence Senior Tech Media
• 1:10
  Speaker 2: Telecom Analyst. He's based in London. So how do you
• 1:14
  Speaker 2: think this is playing out? Matt, as it relates to,
• 1:16
  Speaker 2: we've got some of the leaders in AI saying, okay,
• 1:20
  Speaker 2: we think we need to just kind of slow down
• 1:22
  Speaker 2: and make sure we know what's going on here. Whereas
• 1:25
  Speaker 2: maybe there's some political pressure to keep moving forward. How
• 1:27
  Speaker 2: do you think this might play out?
• 1:30
  Speaker 3: Yeah, I would think in the short term, the best
• 1:35
  Speaker 3: we can hope for in terms of clarity would be
• 1:38
  Speaker 3: some kind of voluntary industry guidelines. It kind of feels
• 1:44
  Speaker 3: to me that we're heading in that direction, given that
• 1:47
  Speaker 3: letter that came out. Over the weekend was kind of
• 1:50
  Speaker 3: endorsed by all the kind of three big main players.
• 1:53
  Speaker 3: So that would be the quickest route to set something
• 1:56
  Speaker 3: that they're all prepared to kind of abide by. And
• 1:59
  Speaker 3: then eventually maybe the kind of U.S. government would. agree
• 2:06
  Speaker 3: on some kind of legislation that would set some more
• 2:08
  Speaker 3: specific or tighter constraints around what isn't permissible. But as
• 2:14
  Speaker 3: we all know, legislation can take a long time to
• 2:18
  Speaker 3: get through the process and doesn't always necessarily end up
• 2:24
  Speaker 3: in a good place in terms of supporting the industry.
• 2:29
  Speaker 3: I think for now... industry guidelines would be the best
• 2:32
  Speaker 3: way forward.
• 2:33
  Speaker 4: That was a very polite way of putting it in
• 2:34
  Speaker 4: terms of Congress getting anything done. Matt, in terms of
• 2:38
  Speaker 4: the IPOs that we're anticipating from OpenAI and Anthropic, OpenAI
• 2:43
  Speaker 4: has made clear it's a 2027 event, not a 2026 event.
• 2:47
  Speaker 4: How are we thinking about what this means for Anthropic
• 2:50
  Speaker 4: filing for an IPO?
• 2:53
  Speaker 3: Yeah, it's a good question. I mean, also, they've been
• 2:54
  Speaker 3: quite quiet on timing. I would imagine they're still keen
• 2:58
  Speaker 3: to go this year. That's just my view, though. Also,
• 3:03
  Speaker 3: they've kind of been flagging recently through press channels that
• 3:08
  Speaker 3: the company was profitable for a second quarter in a row.
• 3:11
  Speaker 3: So that kind of suggests that they're still kind of
• 3:13
  Speaker 3: maintaining that momentum there. As we all know, IPO schedules
• 3:18
  Speaker 3: are incredibly sensitive to stock market sentiment. I think it
• 3:23
  Speaker 3: will really flip on whether this selling pressure we've seen
• 3:29
  Speaker 3: in the last couple of days persists or whether we
• 3:32
  Speaker 3: get a bit of a bounce back. Obviously, there's a
• 3:34
  Speaker 3: lot of debate out there yesterday and today about the
• 3:37
  Speaker 3: extent to which any guardrails or self-imposed restrictions would slow
• 3:43
  Speaker 3: down demand for AI. And that would obviously impact the
• 3:49
  Speaker 3: revenue outlook for Anthropic. So I think that's going to
• 3:51
  Speaker 3: be quite critical. My own perspective on that is that
• 3:54
  Speaker 3: I don't see any near-term slowdown in demand. The models
• 4:00
  Speaker 3: that enterprises have at their disposal today are plenty sophisticated
• 4:05
  Speaker 3: enough for them to deploy very sophisticated overlays to their businesses.
• 4:14
  Speaker 3: I think that the demand outlook is still going to
• 4:16
  Speaker 3: look pretty robust.
• 4:19
  Speaker 2: We had Dan Ives on earlier in the surveillance program
• 4:22
  Speaker 2: this morning, and not surprisingly, he said the money's going
• 4:26
  Speaker 2: to continue to flow, the investment's going to continue to,
• 4:29
  Speaker 2: the companies are going to continue to invest in AI,
• 4:32
  Speaker 2: no slowdown from that perspective. Is that kind of what
• 4:35
  Speaker 2: we're hearing from the companies themselves?
• 4:39
  Speaker 3: Yeah, it's still early days, I think, but I think
• 4:42
  Speaker 3: that's true, United. And I think to some degree, you know,
• 4:45
  Speaker 3: if there is some agreed set of principles about a
• 4:48
  Speaker 3: slowdown in, you know, the kind of real leading edge
• 4:51
  Speaker 3: frontier model development, a lot of that investment is going
• 4:55
  Speaker 3: to find its way into other parts of their existing
• 4:58
  Speaker 3: models that they can make more competitive to kind of
• 5:02
  Speaker 3: create differentiation or perhaps kind of look in other avenues.
• 5:06
  Speaker 3: So rather than kind of pushing forward the boundaries of
• 5:12
  Speaker 3: artificial intelligence. They start to look at the boundaries around
• 5:15
  Speaker 3: efficiency of these models in terms of power consumption or
• 5:19
  Speaker 3: speed or other parameters. And so I think there's still
• 5:22
  Speaker 3: lots of areas that that investment can go to.
• 5:24
  Speaker 4: Matt, what will you be listening for and watching for
• 5:28
  Speaker 4: next as the public outcry over AI safety concerns continues
• 5:33
  Speaker 4: to build?
• 5:35
  Speaker 3: Yeah, personally, I'm kind of going to be primarily focused
• 5:39
  Speaker 3: on what the companies say. My view would still be
• 5:43
  Speaker 3: that the next or kind of the near term goal
• 5:47
  Speaker 3: would be to create some kind of. voluntary industry framework
• 5:52
  Speaker 3: that they'll agree to abide by. But obviously, you know,
• 5:56
  Speaker 3: there's going to be an enormous amount of political rhetoric
• 6:01
  Speaker 3: from every direction about what should or shouldn't be done.
• 6:05
  Speaker 3: But as I said, you know, I don't think anybody's
• 6:07
  Speaker 3: expecting any legislation to kind of be imposed anytime soon.
• 6:12
  Speaker 3: And my sense would be from everything we've seen through
• 6:14
  Speaker 3: the history of the U.S., whatever legislation comes is going
• 6:18
  Speaker 3: to be fairly light touch.
• 6:20
  Speaker 2: So here, Matt, as you're probably aware, AI is getting
• 6:24
  Speaker 2: some pushback from just the marketplace in general, from consumer groups.
• 6:29
  Speaker 2: It's become political here. How is it in the UK
• 6:33
  Speaker 2: and Europe? How's the AI vibe amongst consumers?
• 6:39
  Speaker 3: Yeah, I think it's holding up pretty well. I mean,
• 6:42
  Speaker 3: I'd say that probably relative to the US, the European markets,
• 6:47
  Speaker 3: there was perhaps a little bit more kind of So
• 6:50
  Speaker 3: they're less gung-ho about kind of adoption. Adoption is pretty good. But,
• 6:54
  Speaker 3: you know, I think perhaps that there's already more checks
• 7:01
  Speaker 3: and balances in most European markets about, for example, the
• 7:05
  Speaker 3: deployment market. of new data centers, the environmental impacts that
• 7:09
  Speaker 3: could have. So I think the kind of pace of
• 7:12
  Speaker 3: development along the lines of those kind of perhaps environmentally
• 7:16
  Speaker 3: or politically sensitive dimensions is kind of slower. So things
• 7:23
  Speaker 3: are not quite as magnified here as they are in
• 7:26
  Speaker 3: the US, I wouldn't think.
• 7:29
  Speaker 1: This is the Bloomberg Intelligence Podcast, available on Apple, Spotify,
• 7:34
  Speaker 1: and anywhere else you get your podcasts. Listen live each weekday,
• 7:38
  Speaker 1: 10 a.m. to noon Eastern on Bloomberg.com, the iHeartRadio app, TuneIn,
• 7:43
  Speaker 1: and the Bloomberg Business app. You can also watch us
• 7:46
  Speaker 1: live every weekday on YouTube and always on the Bloomberg Terminal.