Marketplace Tech With Molly Wood

The ongoing government shutdown has caused a lapse in the Cybersecurity Information Sharing Act,a law that's key to protecting the nation against major cyberattacks. Plus, OpenAI this week made a huge investment in chipmaker AMD, the latest in a round of blockbuster deals. And Google said it will be updating its smart home devices with its advanced artificial intelligence, Gemini. Marketplace’s Nova Safo spoke with Maria Curi, tech policy reporter at Axios, about all these headlines on this week’s Marketplace Tech Bytes: Week in Review.

A recent study of nearly 20,000 University of California, San Diego Health workers found cybersecurity training reduced the likelihood of successful phising attacks by just 2%. Marketplace’s Nova Safo spoke with Ariana Mirian, senior researcher at the cybersecurity firm Censys and co-author of the study, who explained that many workers are just not taking those training programs seriously enough.

AI slop is pointless content shared online like fake images and videos. A new study in the Harvard Business Review wants us to consider a variant: “workslop,” or AI-generated reports, emails and more that are sloppily crafted.The authors of the study say 40% of workers they surveyed have encountered workslop. And that's costing time and money. Marketplace’s Nova Safo spoke with social psychologist and vice president of BetterUp Labs Kate Niederhoffer, who co-authored the study. She said workslop harms interpersonal work relationships, such as when one employee receives an AI written report from another.

The number of scams involving job postings and offers nearly tripled between 2020 and 2024, according to the Federal Trade Commission. The change translates to more than $500 million in reported losses. These scams have proliferated on legitimate job boards and you might have even received some via email or text message — fake recruiters reach out with “enticing” offers which are too good to be true. Marketplace’s Nova Safo spoke with Mark Anthony Dyson, author of the Job Scam Report newsletter, about how to spot scams.

Algorithms, which are just sets of instructions expressed in code, are harder to restrict than physical goods. But, as Marketplace’s Meghan McCarty Carino reports, governments — including the U.S. — have long tried to prevent their export.

California passed a sweeping law setting up new AI safety rules this week. Meanwhile, YouTube settled a lawsuit brought by President Trump over account suspensions in the wake of the January 6 capitol riot. And an AI-generated “actor” stirred up controversy in Hollywood and pretty much everywhere else. Marketplace’s Nova Safo spoke with Natasha Mascarenhas, reporter at The Information, to learn more about all these stories on this week’s Marketplace Tech Bytes: Week in Review.

So far this year, LA defense tech companies have raised $4 billion in funding — more than double last year's haul, according to Los Angeles Times tech reporter Queenie Wong. She told Marketplace’s Nova Safo that venture capital firms are increasingly embracing military-focused tech.

The role of artificial intelligence in mental health care is an unsettled issue. States including Illinois, Utah, and Nevada limit or ban the use of AI for therapy. And researchers say such conversations can sometimes veer off course and even be dangerous. Marketplace’s Nova Safo spoke with Jenna Glover, chief clinical officer at the mental health care platform Headspace, which launched an AI assistant, Ebb, last year.

Colleges and universities are dealing with a growing problem of college admission essays being written by artificial intelligence. Rather than try to figure out which essays are AI-generated, some schools are turning to an alternative approach: Have students submit a video instead. Marketplace’s Nova Safo spoke with Anne Kim, senior editor at Washington Monthly magazine who recently wrote about this new strategy that's gaining steam in higher education.

Researchers at several universities tested how successful artificial intelligence can be at political persuasion, and found some AI chatbots were 40-50% more successful than a static message at getting people to change their views. And those views often stayed changed weeks later.Marketplace’s Nova Safo spoke with David Rand, one of the researchers involved in the study who’s also a professor of information science and marketing management at Cornell University.