Marketplace Tech With Molly Wood

This week: a report from venture capital firm SignalFire seems to show that despite all its problems, San Francisco is still the place to be for tech startups in the artificial intelligence space. Plus, why Meta is scrapping plans for a superpremium mixed-reality headset and aiming for a lite version instead. But first, the arrest of Telegram CEO Pavel Durov in France has sent shock waves through the tech world. Durov is facing a number of criminal charges. French authorities allege he is liable for illicit activities conducted on the encrypted messaging platform, including child sex abuse and drug trafficking, essentially because of a failure to moderate content. The case highlights longstanding tensions in the tech world between public safety and free speech. Marketplace’s Meghan McCarty Carino is joined by Natasha Mascarenhas, reporter at The Information, for her take on this week’s tech news.

U.S. intelligence officials have confirmed that Iran was behind the recent cyberattack on former President Donald Trump’s election campaign. Using an approach called spear phishing, hackers sent personalized emails to campaign staff containing malware that allowed them to access private information and then leak it. Déjà vu, right? Javed Ali, a former senior counterterrorism official and a professor of practice at the University of Michigan, says Russia created the blueprint for this kind of attack. Marketplace’s Meghan McCarty Carino asked him for his reaction to Iran adopting the strategy.

By now you’ve probably heard that generative artificial intelligence has the potential to supercharge the spread of disinformation in this election year. But with 68 days until Election Day, we haven’t seen the kind of widespread AI misinformation campaigns that experts warned about. Instead, as Will Oremus pointed out in a recent analysis for The Washington Post, we’ve seen a whole lot of silly AI-generated memes. He told Marketplace’s Meghan McCarty Carino that the most recent examples are coming from one particular presidential candidate.

The California Age-Appropriate Design Code Act, passed in 2022, would be among the most sweeping pieces of legislation to protect kids from online harms — if it hadn’t become tangled up in court. The law has two basic requirements: first, that tech companies analyze and report on whether their products are harmful for children; second, that they minimize how much data they collect from those under 18. Earlier this month a federal appeals court found that first part likely violates the First Amendment, and upheld a lower-court decision blocking that part of the law. But it vacated an injunction on the second component, the part dealing with data privacy. The decision could point a way forward for similar laws, many of which have also run into legal challenges, Aaron Mackey, free speech and transparency litigation director at the Electronic Frontier Foundation, told Marketplace’s Meghan McCarty Carino.

The social media platform X recently launched a new artificial intelligence feature for premium users: Grok-2, an AI model that can also generate images. And the outputs are a bit less censored than you might see with other similar tools. Experimenters online have been able to generate images of Donald Trump and Kamala Harris brandishing guns, Mickey Mouse smoking a cigarette and some far more disturbing tableaus. Grok claims to avoid images that are pornographic, excessively violent or intended to deceive and added it’s cautious about representing content that might infringe on existing copyright. But the guardrails certainly seem to be on the looser side, in keeping with owner Elon Musk’s hands-off approach to content moderation. Marketplace’s Meghan McCarty Carino spoke with Adi Robertson, senior tech and policy editor at the Verge, about Grok-2 and what she found while she tested the AI’s limits.

On today’s show: Google deals with another legal headache. A federal appeals court revived a class-action lawsuit that had been dismissed concerning privacy violations by its Chrome browser. Plus, OpenAI, the maker of ChatGPT, has partnered up with another media brand, Condé Nast. But first, we can’t ignore the biggest happening of the week — the Democratic National Convention in Chicago, where the party formally nominated Vice President Kamala Harris for president with the help of big names in entertainment. Rapper Lil Jon revving up the presentation of the Georgia delegation’s votes was just one viral moment that came out of the convention. What didn’t come out of the gathering, though, were clear indications of the Harris campaign’s tech policy platform. Marketplace’s Meghan McCarty Carino spoke with Maria Curi, tech policy reporter at Axios, about what that platform might look like.

New faces are mingling among the party faithful and the swarm of journalists at the Democratic National Convention in Chicago this week. In a bid to reach younger, more online voters, the DNC invited 200 content creators to cover the convention. One of them is Malynda Hale, who’s been sharing her experience with her more than 50,000 followers on Instagram. Creators like her, she said, have their own part to play at the event. Marketplace’s Meghan McCarty Carino spoke to Hale how it’s been going at the convention.

Earlier this year, the U.S. surgeon general called for a warning label minors on social media, and a growing number of states are requiring online age verification for certain sites. Now, many platforms are adding a safeguard that comes with its own set of trade-offs: facial scanning. These systems use artificial intelligence to analyze visual clues, sometimes in conjunction with a government-issued ID, to keep those it deems too young from accessing a site. In the process, they collect all kinds of identifying data, and like any tool, these are susceptible to errors. In this case, errors could potentially bar adults from parts of the internet. Tech reporter Drew Harwell recently wrote about the proliferation of these systems and the risks that come with them for The Washington Post. He told Marketplace’s Meghan McCarty Carino that they’re getting hard to avoid for internet users.

On July 23, 1999, the Chandra X-Ray Telescope was launched into space aboard the shuttle Columbia, the first such mission to be commanded by a woman astronaut. The Chandra telescope differs from the Hubble telescope, which observes visible light. Or, the James Webb telescope, which captures frequencies in the infrared range. Chandra detects high energy X-rays. But NASA budget constraints could leave the mission going dark in coming years. To learn more about the Chandra Observatory, Marketplace’s Meghan McCarty Carino called its program manager Megan Lin.

When Apple introduced its Screen Time report, it promised the feature would  empower users to manage their device time and balance the things that are really important. But is it actually doing that? Caroline Mimbs Nyce, a staff writer at The Atlantic, recently wrote about why she thinks Screen Time is the worst feature Apple has ever made. She told Marketplace’s Meghan McCarty Carino that it sometimes feels like Screen Time is doing more guilt-tripping than empowering these days.

It’s Friday which means it’s time for Marketplace Tech Bytes: Week in Review. On today’s show, Instagram is known for pretty pictures but a new report shows it lets a lot of ugly and abusive comments remain on the platform. Plus, Door Dash has dominated the food delivery race. Now it looks like Uber may be catching up. But first, the social media platform X relies on advertisers, so why is it going to war with them? Marketplace’s Meghan McCarty Carino discusses all of this week’s biggest tech stories with Anita Ramaswamy, columnist at The Information.  

Last month, senators from both sides of the aisle formally introduced what could become the first U.S. federal law regulating deepfakes. It’s called the No Fakes Act — short for the Nurture Originals, Foster Art, and Keep Entertainment Safe Act. It comes on the heels of controversies like the one involving the OpenAI voice assistant, which may or may not have sounded a little too much like actress Scarlett Johansson. GOP Sen. Marsha Blackburn of Tennessee helped draft the bill, saying: “The No Fakes Act is vital for these entertainers. It protects their name, image, likeness, their vocal recordings.” The legislation has been applauded by many in creative industries, said Moiya McTier, senior adviser at the Human Artistry Campaign, a group advocating for stronger guardrails on artificial intelligence.

Generative artificial intelligence has made it possible to mimic someone’s voice and generate a script for that voice in real time. The tech, of course, is already used to scam and defraud people, but what if you just had it make a bunch of calls on your behalf? That’s what journalist Evan Ratliff did for his new podcast, “Shell Game.” He trained AI audio clones, gave them phone numbers and sat back as they took on customer service agents, family members, therapists and even a few scammers. Marketplace’s Meghan McCarty Carino spoke with Ratliff — and briefly with one of his AI agents — about his takeaways from producing the show and whether the clones succeeded in tricking people into thinking they were who they said they were.

Susan Wojcicki, the former CEO of YouTube, died last week at the age of 56. As a kid, she wanted to be an artist. As an adult, she discovered the “art of technology.”

Taylor Swift’s “Eras Tour” has shattered records, crashed websites and broken the hearts of fans vying for tickets. The BBC’s Sam Gruet says AI could help fans purchase tickets for a fair price and without the risk of fraud.

On today’s show: Olympians have taken to social media to celebrate, sometimes to trash talk, but also to discuss their mental health. And Google pulled a controversial Olympics ad featuring its Gemini artificial intelligence tool. But first, what the stock market sell-off could be saying about the AI boom. Marketplace’s Meghan McCarty Carino discusses all of this with Christina Farr, author of the health tech newsletter “Second Opinion,” who says there’s growing chatter that AI has gotten a bit overinflated.

Who will win the election? What will the vote margin be? Will Donald Trump post on X before November? People can place bets on all these real-world questions — and more — on prediction markets. And these online platforms like PredictIt and Polymarket are increasingly being looked to as crystal balls in this chaotic election, promising real-time political insights and the chance to make a few bucks. Marketplace’s Meghan McCarty Carino spoke to Chris Cohen, the deputy site editor of GQ, who recently wrote about his experience getting in on the action of what appears to be a prediction market “gold rush.”

More than 20 years ago, executives at rival auction houses Sotheby’s and Christie’s were found guilty of coordinating a massive price-fixing scheme. Leaders from the companies held covert meetings, where they set identical commission fees. Today, active antitrust cases show that the ways in which companies might conspire are changing. Algorithms can replace secret meetings, but U.S. regulators say it’s still collusion, whether it’s a human or a bot pulling the strings. Marketplace’s Lily Jamali spoke to Joe Harrington at the University of Pennsylvania’s Wharton School about how antitrust law holds up against new technology.

In the early days of the COVID-19 pandemic, a photo of two little girls in the parking lot of a California Taco Bell went viral. They were doing their schoolwork on laptops in that inconvenient location because the restaurant provided free Wi-Fi, which they didn’t have at home. The girls came to symbolize the digital underclass that’s emerged since the rise of the internet. There are millions of American kids like them, says Nicol Turner Lee, senior fellow at the Brookings Institution. Her analysis of the digital divide is contained in her new book, “Digitally Invisible: How the Internet Is Creating the New Underclass.”  

Large language models go through a lot of vetting before they’re released to the public. That includes safety tests, bias checks, ethical reviews and more. But what if, hypothetically, a model could dodge a safety question by lying to developers, hiding its real response to a safety test and instead giving the exact response its human handlers are looking for? A recent study shows that advanced LLMs are developing the capacity for deception, and that could bring that hypothetical situation closer to reality. Marketplace’s Lily Jamali speaks with Thilo Hagendorff, a researcher at the University of Stuttgart and the author of the study, about his findings.