St. Louis · Kansas City

Tech Talk Tuesday: AI That Predicts Death From a Photo, Hackers Hijacking Your AI Bill, and More!

An AI model can now estimate a cancer patient’s mortality risk from a single photo of their face, hackers are quietly running up five-figure bills on company AI accounts, and Apple is finally entering the smart home race with a $299 hub. On this week’s Tech Talk Tuesday on the Mark Reardon Show, 97.1 FM Talk, ThrottleNet owner George Rosenthal unpacked all three stories, explained why each one deserves a closer look, and shared one simple Wi-Fi tweak that can keep your smart devices from putting your work laptop at risk. Interested in learning more? Continue reading below.
AI That Reads Your Face to Predict Mortality Risk
The segment opened with a story from The Washington Post that left both George and the host stunned. Researchers at Mass General Brigham, building on an earlier tool called FaceAge that estimates a person’s biological age from a photo, have developed a new foundation model trained on more than 40 million facial photographs. The newest version, called FAHR-FaceSurvival, goes a step further. It can predict a cancer patient’s mortality risk directly from a photograph alone.
George explained that the system analyzes areas such as the bridge of the nose, the area under the eyes, the temples, and the forehead, measuring signs of frailty like soft-tissue loss and muscle wasting. Doctors have long recognized a certain look in patients who are very close to death. What’s new is that AI appears able to spot warning signs much further out. To train it, the researchers drew on photos of more than 30,000 cancer patients whose medical outcomes were already known, including whether they survived and, if not, how long they lived after diagnosis.
George cited a statistic from the coverage: among patients in the highest 10% risk group, about half passed away within 90 days.
Despite the impressive results, George was blunt about his concerns:
- Insurance companies. What happens if insurers get access to these photos and use them to deny claims?
- Lifelong photo collection. What if doctors began photographing patients’ faces at every visit from birth, the way they weigh you every time you walk in?
- False positives and family distress. What happens when a family hears that a loved one may only have months to live and the prediction turns out to be wrong? And could doctors end up performing unnecessary treatments because an algorithm said so?
- Regulation. George called for lawmakers to look at this immediately and wondered whether legislation could realistically keep the technology from being used in harmful ways.
He drew a clear line between this and other medical uses of AI. Having AI analyze a CT scan or MRI is a very different thing, he said, from having it judge your face and make a prediction about your future. That, in his view, is dangerous territory.
Criminals Are Taking Over AI Accounts and Running Up the Tab
George then explained a notable change in how cybercriminals work. In the past, attackers who broke into a network would lock the files and demand payment. These days, before they do that, they search for the API keys, logins, and passwords connected to the AI services a business or its staff rely on, including ChatGPT, Gemini, and Claude.
The motivation comes down to “tokens,” the units of computing power AI providers charge for. Every request, whether it’s a simple question or a task handed to an AI agent, consumes them. A company may pay a few hundred dollars monthly for an enterprise plan, but a criminal with your credentials can burn through tokens nonstop, using your account to go after other companies or to produce malicious code for ransomware and malware attacks. George shared that some businesses have opened a bill showing $25,000 to $40,000 in charges built up over one weekend, when the normal total would have been about $400.
His suggested safeguards:
- Put hard spending limits in place within the settings of every AI service your company uses.
- Enable billing alerts, either through your bank or the AI provider, so you’re warned about your remaining tokens before extra charges kick in.
- Write an AI use policy. If your team is using a mix of AI tools without clear guidelines, you have no visibility into what’s going on. George compared the importance of an AI policy to protecting against ransomware, and said ThrottleNet can walk businesses through creating one.
Apple HomeView: A Smart Home Hub That Skips the Ads
For a lighter topic, George turned to Apple’s new HomeView, the company’s long-awaited answer to the Amazon Echo and Google Home. It’s a $299 smart display that can sit on a counter or mount on a wall like a thermostat, serving as the central hub for your home. Pre-orders are open now, and it ships November 13.
What impressed George most is what it won’t do. Unlike Amazon and Google devices, which regularly push shopping suggestions and products, HomeView is designed to be ad-free. He also pointed to Apple’s emphasis on privacy: the device is built around a rebuilt, AI-powered version of Siri, and Apple says requests are handled on the device itself or through its Private Cloud Compute system. George said you’ll be able to do the same things you can with an Echo or Google Home, such as turning on the TV or the lights with a voice command, while the AI learns your habits over time.
If you’re thinking about buying one, George offered a few notes:
- Look for Matter compatibility. Matter is a common language that lets smart devices from different brands communicate. Lights, thermostats, and other gadgets that don’t support it may need to be replaced.
- Check your existing ecosystem. George has Ring cameras, which are owned by Amazon, and says getting them to work with Apple’s hub may require an extra bridge device costing a couple hundred dollars.
An Easy Wi-Fi Change to Shield Your Work Laptop
George saved his most useful advice for last: move your smart devices onto a guest Wi-Fi network. Cameras, thermostats, refrigerators, TVs, and similar connected products should stay off the network your work laptop uses.
The logic is straightforward. If a hacker gets into one of those gadgets, they won’t have an automatic route to the computer you use for your job. It only takes a few minutes: sign in to your router (whether it came from Spectrum, AT&T, or a store), set up a guest network with a separate password, and switch your smart devices to it.
Both hosts confessed they didn’t know this was possible, which is exactly why George suggests looking into it today.
The Bottom Line
AI is advancing into medicine, business operations, and our living rooms, and each step brings real benefits along with real risks. Set spending caps on your AI accounts, put an AI policy in place at work, keep your smart devices on a separate network, and think hard about how personal data, including a photo of your face, might be used.
Catch George Rosenthal live every Tuesday for Tech Talk Tuesday on the Mark Reardon Show, 97.1 FM Talk. Have a tech question you’d like answered on air? Reach out to [email protected] or contact us here!
