The delusional side of AI therapy.

Please enjoy this encore of Hacking Humans. This week, our hosts⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠Dave Bittner⁠⁠⁠⁠⁠⁠, ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Joe Carrigan⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠, and ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Maria Varmazis⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ (also host of the ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠T-Minus⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Space Daily show) are back sharing the latest in social engineering scams, phishing schemes, and criminal exploits that are making headlines. We've got some follow-up from listener Kajetan, who recalled a run-in with a scammer in Paris posing as a mute fundraiser—and says he performed a "miracle" by crossing out his name, prompting the supposedly mute woman to suddenly start yelling at him. Maria has the story on how small businesses in Toronto, like the family-run Souvlaki Hut and Pippins Tea Company, were shocked to discover that thieves exploited vulnerabilities in their point of sale terminals to issue themselves thousands in fraudulent refunds—exposing serious flaws in how these machines are secured. Dave's story is on a Stanford-led study that found popular AI therapy bots, including ChatGPT and commercial mental health platforms, often respond inappropriately to serious mental health issues—fueling delusions, validating harmful thoughts, and failing to follow basic therapeutic guidelines—raising urgent concerns about their use as replacements for human therapists. Joe follows the story on a sweeping federal investigation into Minnesota's Housing Stabilization Services program, where agents raided homes and businesses tied to an alleged multi-million-dollar Medicaid fraud scheme that exploited vulnerable residents and billed taxpayers for housing support services that were never provided. Our catch of the day is on a patient scammer who spent five months building trust before claiming to send a $700K inheritance payout locked in a lawsuit—complete with a fake video of a safe and a shady tracking number—only to demand €15,000 in "customs fees," a scam the Redditor thankfully saw through before handing over any money. Complete our annual ⁠⁠⁠⁠audience survey⁠⁠⁠⁠ before August 31. Resources and links to stories: ⁠⁠AI therapy bots fuel delusions and give dangerous advice, Stanford study finds⁠ ⁠‘It was a shock’: Toronto business owner says customer used point of sale terminal to issue himself $2,000 refund⁠ ⁠KARE 11 Investigates: Federal agents raid homes & businesses seizing evidence in housing fraud investigation⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Have a Catch of the Day you'd like to share? Email it to us at ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠hackinghumans@n2k.com⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠.