Marketplace Tech

OpenAI's new teen-appropriate version of ChatGPT may not be living up to its promises, according to an independent evaluation. The new mode is supposed to have stronger guardrails for sensitive topics, automatic parental alerts and a study tool that won't just give kids the answers to their homework.

But the nonprofit Common Sense Media evaluated the tool and found it poses "unacceptable risks." Tom Siegel leads the Youth Safety Institute at Common Sense and tested about 4,000 prompts using fake teen accounts. He spoke with Meghan McCarty Carino about his findings.

More on this:

“ChatGPT for Teens Poses Unacceptable Risk to Kids, Common Sense Media Finds” from Common Sense Media

This story includes references to suicide, self-harm and mental health crises. If you or someone you know is experiencing a mental health crisis, call or text 988 to reach the Suicide & Crisis Lifeline. Available 24/7.

A growing number of lawsuits accuse AI companies of harming users — alleging chatbots have encouraged self-harm, reinforced delusions, or failed to recognize signs of distress.

And yet at the same time, a growing number of users, particularly younger ones, say they're turning to AI for emotional support.

Dr. John Torous studies how technology interacts with mental health. He says in the case of AI, it's complicated.

More on this:

Nearly 2 out of 3 teens and tweens use AI chatbots, new research shows — from NPR

An independent benchmark for AI in mental health — from MindBench.ai

Development, validation and AI chatbot evaluation of the Adolescent Mental Health AI Response Tool — from Nature

Interpol estimates scam losses totaled almost $500 billion globally last year.

In many cases, it’s not just one bad actor but an entire network of scammers. And Cambodia has become a major hub for the industry. Last year, the U.S. issued more than 100 sanctions tied to fraud operations in the country.

Reporter Leila Goldstein takes us behind the scenes of an abandoned scam compound in southeast Cambodia.

In September, OpenAI said that its models had solved one of math's famous Millennium Prize problems. Then, the lab dropped more than 700 papers tackling another 400 complex problems with more breakthroughs to come.

These developments have caused something of an existential crisis in the math world. But Alex Kontorovich, a professor of math at Rutgers University, says they're a gift.

More on this:

“Sharing AI progress in mathematics” from OpenAI

“OpenAI Releases Findings on 377 Math Problems, Further Roiling Field” from The New York Times

“The Shape of Math To Come” from arXiv

Anthropic formally filed to go public this week and about a third of its paperwork is dedicated to potential risks — including political ones. We'll get into it on today's “Tech Bytes: Week in Review.”

Plus, it’s a big week for big tech earnings: Microsoft pleased investors with a more restrained approach to AI spending, but still announced it’s offering buyouts to 60,000 employees. Meanwhile, Google’s parent company Alphabet took a plunge on Wall Street.

Here to help us unpack this week’s tech news is Maria Curi, tech policy reporter at Axios.

Check out our YouTube page to watch more episodes of “Tech Bytes.”

More on this:

Anthropic warns investors of AI’s ‘existential risk to humanity’ in IPO prospectus, reports say — from CNBC

Microsoft is offering buyouts to around 60,000 employees. Read the memo — from Business Insider

Meta eyes record $97.5 billion equity raise to fund AI expansion — from Reuters

Scoop: FTC demands Anthropic execs testify in probe — from Axios

Consumer electronics and AI data centers need memory chips. But there are only a handful of companies in the world that make them. The surge in demand has caused memory shortages and price spikes for everything, including our all-important smartphones.

Gerrit Schneemann is principal analyst at Counterpoint Research, which recently reported global phone prices have jumped by an average of 15%, and even more in markets where consumers can least afford it.

More on this:

“The AI boom is making the world’s cheapest smartphones disappear” from Rest of World

“Global Existing Smartphone Prices Climb About 15% in 2026, New Launches 25% Costlier YoY as OEMs Maneuver Memory Price Hikes” from Counterpoint Research

Halloween haunts have long relied on a little technical wizardry to bring their terrors to life —from lighting and animatronics to digital decorations.

But these days, the technology behind those scares is getting increasingly sophisticated, and the spectacles more elaborate.

We got a glimpse behind the scenes at the Hollywood Haunted Hayride, put on by Thirteenth Floor Entertainment in Los Angeles.

Listen for more … if you dare.

Urgent concerns about losing control of AI have burst into the national conversation recently after the OpenAI hack of the open-source AI platform Hugging Face, the high-profile resignation of an Anthropic researcher, and other AI insiders sharing alarmingly high p(doom) figures — that is, the probability that AI kills us all.

An event listing for a picnic beckoned people in the AI scene to spend a Sunday afternoon talking about recent calls to slow down the technology.

Dony Christie, a community organizer in the AI safety scene, chose this spot not for the scenery but because it’s between the offices of Anthropic and OpenAI.

“The hope is, why can't people at the labs just meet at a park in between them and work out a deal?” he said. “I have been concerned about existential risks for a long time. Just like, what do I do to save the world?”

Donna Guthrie had obituaries on her mind after her brother passed away in May. She recently celebrated her 80th birthday and started wondering how she might be remembered. So, she asked ChatGPT to write her obituary. It's the topic of today's “Uncanny AI,” our regular segment about our sometimes unsettling experiences with this tech.

Plus, we asked Michelle Huang, a technologist and artist who explores the use of AI for wonder, if this is creepy or healthy.

More on this:

“How AI can connect you with your inner child” from Marketplace Tech

Apple's long-awaited iPhone Duo launched this week, and everyone has the same question: can you see a crease? Joanna Stern, founder of media company New Things, has been putting it to the test for her newsletter. She's tried apps, phone calls, cameras. But, she really wanted to know if the much-hyped folding phone would stand the test of time. She hired a hardware designer to build a machine that opened and closed the iPhone Duo over and over while she slept — more than 32,000 times. Stern joins joins Marketplace’s Meghan McCarty Carino on this week’s “Tech Bytes: Week in Review.”

Plus, Elon Musk says there could be a merger between SpaceX and Tesla. And, recent court documents reveal internal warnings about child safety issues were ignored by TikTok executives.

More on this:

“iPhone Duo Review: The Good, The Bad, and The Buggy” from New Things

“Apple announces product launch event for Oct. 28. Analysts expect touchscreen laptop” from CNBC

“SpaceX and Tesla Are Already Joined at the Hip. What a Formal Merger Would Mean.” from The Wall Street Journal

“Musk Ordered Starlink Slowdown That Blindsided Ukraine” from The Wall Street Journal

“TikTok staff raised child safety concerns, but executives reportedly failed to act” from TechCrunch