Skip to content
Meta CEO Weighs In On AI Safety Debate artwork

Meta CEO Weighs In On AI Safety Debate

Bloomberg Tech

September 16, 2026

Bloomberg’s Ed Ludlow breaks down Mark Zuckerberg's call for AI labs to rely on independent evaluators and advisers to ensure models are safe. Plus, OpenAI is said to be in early talks with investors about a new funding round that could value the ChatGPT maker at more than $1.

Speakers Ed Ludlow, Riley Griffin, Shireen Ghaffari, Summer Mersinger, Tom Mueller, Natasha Mascarenes, Maggie Eastland, Peggy Johnson, Slater Stitch, Yajaira Anand

TopicsTech NewsNewsBusiness News

SPEAKER_1 (0:02)

Bloomberg Audio Studios, Podcasts, Radio, News. Bloomberg Tech is live from the heart of Silicon Valley with Ed Ludlow in San Francisco.

Ed Ludlow (0:22)

This is Bloomberg Tech, coming up, Mark Zuckerberg weighs in on the global AI safety debate and says AI labs should rely on independent evaluators and advisors to ensure models are safe. Plus, OpenAI could be looking to raise more money before it heads to the public markets, and a new funding round could value it at more than $1.2 trillion. And Impulse Space announces a $308 million Series D extension. We speak with the CEO later this hour. Our top story, Mark Zuckerberg has weighed in on the AI safety debate. The Meta CEO said AI labs should rely on independent evaluators and advisors to ensure that their models are safe, rather than forcing an industry-wide slowdown. He wrote on X that the company had delayed the release of its latest AI tool, Mules, for several months to ensure it was safe. Bloomberg's Riley Griffin, who leads our coverage of Meta, is with us. Very simply, Mark Zuckerberg is saying, this is for each lab to do itself. Just get on with it.

Go through the post.

Riley Griffin (1:22)

Yeah. I mean, the most notable thing about the post to me is he's throwing a little bit of shade at his competitors. Here with the tweet that we have on the screen, he effectively has said, we did a slowdown. We did it internally. We didn't raise a kerfuffle about it. And we didn't call on our peers to slow down too. So Meta is moving ahead. It's calling on independent evaluation. But it's not saying, hey, right now, we're going to slow down after these calls.

Ed Ludlow (1:49)

You know, some like camps are emerging, right? So, you know, Dario Amodei in his Anthropics SA on Saturday explicitly said, we think independent evaluation is important.

But where, you know, Zuckerberg disagrees with Amodei is, there doesn't need to be this coalition of the frontier labs together. You talked about safety at Meta with Dina Palma Cormac, right? Explain who she is, but also her answer to your question, if there was one.

Riley Griffin (2:14)

Yeah, so on Monday, I spoke with Dina Palma Cormac in Meta's headquarters in Menlo Park. She is the president of the company, someone who came in just earlier this year from the board.

She is playing a really broad and sweeping role with her hands all over the AI story. And in this case, she said that safety should be baked in from the bottom up. It should be a part of the process day to day. Again, it's a polite way of saying, we're not going to commit to slowing down right now. We think this is a part of our process.

Ed Ludlow (2:45)

She is in the Jensen 1 camp, which is safety is an engineering problem.

It's for the companies themselves to say, if we're good at AI, we should make sure it's safe. That conversation with Palma Cormac is interesting. The question I ask myself is, is Dina Palma Cormac the present day Sheryl Sandberg? How does Mark Zuckerberg rely on her as the face of the company?

Riley Griffin (3:09)

I think it's a comparison that's made often because of their backgrounds in Washington, their connections to the financial sector, their focus on mentorship.

Ed Ludlow (3:18)

She was a Trump administration official. She was at Goldman Sachs.

Riley Griffin (3:20)

Correct. And Sheryl Sandberg notably had more connections to the Democratic Party. But in this case, there's a big difference. Sheryl grew the company from something small to what it is now.

Dina is coming into a $1.7 trillion company trying to smooth out its rough edges. They have a lot of challenges ahead, particularly when it comes to the AI backlash, focus on data centers. And she is both helping to clinch deals and retell that AI story.

Ed Ludlow (3:49)

So those deals and that focus on data centers, Meta is massively committed to spending on compute. And reading your reporting and the statements from Meta, Dina Palma Cormack is kind of overseeing all facets of that, right? It's a really big portfolio for her.

Riley Griffin (4:05)

She's everywhere all at once. And not only is she working on compute, I mean, something that emerged, she was critical to the up to $18 billion settlement with the state attorneys general. That is altogether separate from the compute story, but she is a right hand to Mark Zuckerberg. She has really earned his trust.

She left our interview to go have dinner at his house after four hours of conversation that very day. So she's very involved. CJ Mahoney, who is the chief legal officer, said she's the last person consulted on every decision. So it's notable that she has risen to this level in such a short amount of time.

33 more minutes of transcript below

Thousands of transcripts fetched by people building searchable podcast archives

Fetch the whole transcript

The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000790222984