Rohit Chopra’s AI nightmare; Meta faces new trial on social media addiction in LA

SPEAKER_1 (0:00)

Flowing ad budget on metrics that look great till the CFO sees them, that's bull spend. And marketers are calling it out in Dashboard Confessions.

SPEAKER_2 (0:09)

I remember telling my boss, it'll be good for the brand when leads were slow. Yeah, it wasn't.

SPEAKER_1 (0:17)

Cut the bull spend. LinkedIn lets you target by company, job title and more. Advertise on LinkedIn.

Spend $250 on your first campaign and get a $250 credit. Go to linkedin.com/campaignterms and conditions apply.

Sam Sutton (0:36)

Hello, and welcome to Decoded by POLITICO. I'm Sam Sutton. Today on the show, my conversation with Rohit Chopra, the former leader of Washington's primary consumer watchdog, who's now working as one of Governor Gavin Newsom's top regulators. We spoke at length about the risks he's seen around the AI investment boom, the use of agents, and his own future as Democrats ready for tech and financial services agendas for after the midterms. Plus, POLITICO reporter Tyler Katzenberger joins me to talk about the upcoming trial against Meta and SNAP.

We're going to start today's show with my conversation with Rohit Chopra. I wanted to talk to Chopra for a couple of reasons. He's been a major player in progressive circles since the financial crisis. He was an Elizabeth Warren protege and former FTC commissioner, who went on to be the director of the Consumer Financial Protection Bureau under President Joe Biden, and now he's on Team Newsome. He's the secretary of the new California Business and Consumer Services Agency. It's a newly created role that gives him oversight of everything from the state's banking sector and its medical board to cannabis sales and horse races. And it provides him a really interesting purge to be a power player in state level oversight of AI.

But I also wanted to know how he felt about AI's broader role in the economy, the investment boom, and how he might approach regulating the use of AI by companies that are operated and licensed in California. Here's a section of our conversation.

Secretary Chopra, thank you so much for joining me.

Rohit Chopra (2:00)

Sam, thanks for having me.

Sam Sutton (2:02)

I want to start by talking about agentic AI agents in banking. More tech companies have launched AI agents that interact with folks' bank accounts or their investment portfolios.

How are you thinking about the scope of risk to consumers against that backdrop?

Rohit Chopra (2:16)

Well, Sam, AI agents really have their predecessors in so many different types of forms over the years.

And when we step away from banking, we saw them on social media accounts, bots that were amplifying certain types of content or putting certain types of comments on posts with a particular strategic purpose, sometimes even geopolitical. So we've seen this in many different formats of ways in which bots can be used to manipulate markets and to sometimes create impressions for consumers that may not be accurate. I saw this pretty close at the Federal Trade Commission when it came to things like fake reviews.

All sorts of consumer issues come up where people are deceived and led to believe one thing when the reality is much different. So now we're talking about agents or bots getting inside the premises of our financial infrastructure. And that actually opens up a whole other set of concerns about cyber security, about fraud, about account takeovers, and also enormous amounts of data harvesting by those agents.

And where are they reporting that information to? Those are some of the key questions that we are working hard in California to figure out and address because it's very clear in Washington, they're pretty much asleep at the switch.

Sam Sutton (4:04)

So how are you thinking about the liability piece of this then? Because as you said, like when you were at the CFPB, it was pretty clear that it was the financial institutions themselves that were rolling out these products. Now we have tech developers and a large number of tech developers that have, you know, agentic products that consumers can use or could be used by bad actors to access financial systems, financial data. Are you, you know, looking at going after the developer specifically here from a liability perspective who may be at fault here?

Rohit Chopra (4:36)

Let's think of all of the parties in a potential problematic transaction and whether that's about a bank or whether it's about some other type of online account you may have.

You may be using some sort of agent that you have given specific directions to, and those directions may be to engage in some sort of commercial transaction, financial transaction. And there's a question about when there is a specific screw up, where you are harmed. To what extent when that agent has deviated from your instructions or engaged in conduct that you would have never reasonably assumed, that's where there needs to be some serious scrutiny of the agent and the company that produced the agent itself. There's separately a question about when is a financial institution or other type of online service provider, when are they accepting instructions that do not reasonably seem to be an instruction of the consumer themselves. That also raises some scrutiny over them. Because Sam, what has happened is that you have seen a whole range of companies trying to drastically reduce their costs by shifting to AI agents.