What We Know About the Engineer Accused of Bugging OpenAI Conference Room

SPEAKER_1 (0:01)

Running a business means keeping pace with the latest tech. You put a lot into building your enterprise software stack, and scrapping it to integrate AI is out of the question.

That's why IBM technology works with all the apps, data and models you need. It's another way IBM helps you build on the technology you rely on. Let's create smarter business. IBM. Visit ibm.com to learn more.

Alex Ossola (0:33)

Silicon Valley was already bracing for honeypots and insider threats before an OpenAI engineer bugged a company conference room. Plus, longtime competitors Meta and Apple are teaming up to bring open-weight AI to Mac computers.

Meghan Bobrowsky (0:47)

AI is a new era for everything. Technology is moving so quickly, the stakes are so high that you might say, hey, it's actually maybe a good idea for me to partner with my rival here.

Alex Ossola (0:59)

And President Trump reverses Defense Secretary Pete Hegseth's decision to livestream the execution of Fort Hood shooter Nidal Hasan. It's Thursday, October 15th. I'm Alex Ossola for the Wall Street Journal. This is the PM Edition of What's News, the top headlines and business stories that move the world today.

We begin our show tonight with more details about an incident that's rattled Silicon Valley. Abdulrahman Alfozan, a 35-year-old engineer for OpenAI and a Saudi national was arrested this week. The Justice Department charged him for violating a federal recording statute. And it's exactly the kind of incident big tech companies have been working hard to try to prevent. Max Zeff covers the AI industry for the journal and joins me now.

Max, Alfozan had been working at OpenAI for about a year before this incident. What do we know about what he did and why he did it?

Max Zeff (1:50)

Most of what we know about what Alfozan did comes in the form of a Department of Justice complaint that was filed in a federal court last week.

In the complaint, they allege that Alfozan effectively scoped out a top secret conference room at OpenAI's office, which he did not have access to. And he inserted a microphone behind a TV and wrapped it in toilet paper. And the allegation is that he wanted to extract the comments and conversations that were being said in that room.

Alex Ossola (2:24)

Alfozan wound up getting caught on camera and arrested after OpenAI found his microphone in a regular security sweep. What does Alfozan or his attorney say about this?

Max Zeff (2:34)

Alfozan was charged with putting a recording device in OpenAI's office. You're not allowed to record people without their consent. But he wasn't charged with being a spy. And Alfozan's lawyer in a statement to us has basically said, we would hope that people wouldn't jump to conclusions here. He was just charged with putting a recording device here. And we're still waiting to learn more.

Alex Ossola (2:55)

Where is he now? And what can we expect to come next in his case?

Max Zeff (2:59)

Alfozan is currently in federal custody. He's expected to appear next week before a federal court. He's expected to have a speedy trial. We'll have to see what else unearths in this case and whether there are any alleged ties to Saudi Arabia. And maybe he can explain himself as to why he put that recording device in OpenAI's office to begin with. For the charge he's facing right now, he could face up to a few years in prison and a $250,000 fine. If new allegations are brought, I mean, that could change. But I mean, also, it's hard to think about what his future job prospects might be after this.

Alex Ossola (3:36)

So, interestingly, just before Alfozan's arrest, OpenAI had held training sessions to prevent exactly this kind of scenario. How common are spying incidents? And are tech companies getting more worried about them now?

Max Zeff (3:50)

These kinds of spying incidents aren't exactly common, but tech companies are warning their employees to be aware of stuff like this. You know, in one example that my colleagues reported on, OpenAI actually tells employees, if you're a young man and you're at a park and a beautiful woman comes up to you to ask yourself, is this a spy trying to get information out of me from where I work? Which is kind of just a hilarious example of just kind of the paranoia at the companies developing frontier AI systems these days.

Alex Ossola (4:21)

What are these organizations or countries that are potentially spying on companies like OpenAI doing it for? What information are they hoping to get?

Max Zeff (4:28)

In these executive conference rooms that Alfozan tapped, it's not like they were reading out, you know, line by line the source code of OpenAI's next model. The conversations that were happening in that room were probably much more high level plans about designs for future AI models, plans for future products, kind of road maps for the business. These are still incredibly sensitive top secret details, but it's possible that just any of this information could help another country or a competitor, anyone else, get a leg up on OpenAI.