
The voice that opened this week’s radio segment sounded exactly like George Rosenthal, but George never said a word of it. It was an AI clone, and it sets the stage for a story about deepfake fraud that has cost companies more than $100 million. On this week’s Tech Talk Tuesday on the Mark Reardon Show, 97.1 FM Talk, ThrottleNet owner George Rosenthal also tackled the nationwide ground stop that halted U.S. flights, the big Microsoft outage, the shake-up at Nvidia, and OpenAI’s new wallet that lets ChatGPT shop for you, plus the guardrails to set before you try it. Interested in learning more? Continue reading below.
That Wasn’t George: Inside the AI Voice Clone
The segment began with a short clip that sounded just like George. In reality, every word was typed, then spoken by an AI clone of his voice. George explained that it took about two hours of reading scripts aloud to train the voice. A rushed first attempt, made in about 10 seconds, sounded stiff and unnatural. But when he took his time, with careful prompts about tone, inflection, and energy, the result was nearly identical to the way he really speaks. His parents couldn’t tell the difference.
His point is that the technology has reached the stage where a cloned voice could leave you a convincing voicemail, or even sit through an entire phone meeting, without you realizing it isn’t the real person.
Deepfake CEO Fraud: A Growing Epidemic
That’s exactly the tactic behind a case the Justice Department just announced. Federal prosecutors in Manhattan have charged three men in a scheme that allegedly used AI to impersonate company CEOs and lawyers and collected about $129 million from businesses between 2024 and 2026. The charges are allegations, and the men are presumed innocent.
According to the reporting, the scammers called employees posing as an executive, using AI to mimic the person’s voice or image, and asked for help with a supposedly secret deal. The fake executive then directed the employee to a “lawyer” at a large firm, played by another member of the group, who supplied bank details for a large transfer. Victims reportedly included a Missouri hospital system, a UK fintech company that lost about $11.6 million, a bank, a mutual fund, and a pet food company.
George pointed out that attackers don’t need much to clone you. A few clips of your face or voice from YouTube, Facebook, or any public video can be enough, and they can feed it into an AI with a bit of background and tell it to carry on a conversation.
His advice for businesses and families:
- Verify before you act. If a voicemail, call, or video chat from your boss asks you to do something, don’t do it until you’ve confirmed. Call them back at a number you already know, walk down to their office, or text their direct number.
- Treat urgency and secrecy as red flags. Phrases like “this is a secret project” or “it has to go through in the next couple of days” are classic pressure tactics.
- Agree on a code word. Have a code word inside your company for sensitive requests, and set up a safe word with your family too. George says it’s a good defense against calls claiming a child has been abducted, since the voice on the line could be cloned. He and his daughter have already done this.
The Nationwide Ground Stop and a Big Microsoft Outage
George and the host then turned to two big disruptions from the past week. On Friday, November 13, the FAA halted all U.S. flights for nearly four hours after a communications failure. Reports say more than 9,000 flights were canceled. Officials have said an upgrade went wrong: contractor crews were reportedly replacing aging copper wire with fiber optic cable when a router began flooding the network with log-on attempts. The transportation secretary has said the investigation is ongoing and has not blamed the contractor.
The day before, a global Microsoft outage knocked out Outlook, Teams, and other services for hours. Microsoft has said that a number of servers became unexpectedly unavailable, causing load to cascade across regions, and that it sees no indication of malicious activity. It hasn’t published a root cause.
George noted that none of the research he’s seen links these events, and nobody has claimed responsibility for any of them. But he pointed out that not being connected to each other doesn’t mean cyberattacks are ruled out, and neither organization has confirmed a cause. On the flight system, he was blunt about what’s behind the fragility. The FAA is relying on aging equipment, and officials have said they’ve had to search eBay for parts that are no longer manufactured. His takeaways:
- Do upgrades off-hours. If critical equipment is being replaced, schedule it for the middle of the night, not the middle of the afternoon.
- Always have a rollback plan. Keep the old equipment ready to put back in place if the new setup fails.
- Make it a priority. Funding for aviation and the power grid should be at the top of the list. George said he’d support the investment as long as it’s transparent, and that this shouldn’t be a partisan issue, since everyone flies.
Nvidia’s CEO Steps Down
George also weighed in on the latest from Nvidia. The company announced on November 12 that Jensen Huang was stepping down as CEO and board member, effective immediately, with CFO Colette Kress named interim CEO. Nvidia’s announcement did not give a reason. Huang has since posted a message thanking supporters and saying he would be honored to return if the board decides it’s in the company’s best interest.
George said he doesn’t expect the board to bring him back. He called Huang’s story, from starting with nothing in 1993 to building one of the most valuable companies on the planet, both a remarkable success and a loss, and said he’s curious who will lead Nvidia next.
ChatGPT’s New Wallet: Convenience With Guardrails
Closing on a technology you may be tempted to try before the holidays, George covered OpenAI’s new ChatGPT Wallet, announced November 13. It works somewhat like Apple Wallet, letting ChatGPT’s AI agents make purchases for you. It’s available in the U.S. for ChatGPT Pro subscribers who are 18 or older. According to reports, the system uses a payment token rather than your real card number, and you approve each purchase with Face ID or a fingerprint.
George is excited about the possibilities and encouraged listeners to try it, but offered these safety tips:
- Use a credit card, not a debit card. You can dispute credit card charges more easily. George’s advice was clear: don’t give AI your debit card.
- Require cart review. Tell the AI you need to see the shopping cart and approve the purchase before it checks out.
- Turn on purchase alerts. Set up text alerts with your card issuer for transactions over a low threshold, such as $20 to $100, so you’ll know right away if the AI buys something unexpected.
George also noted that Amazon is blocking AI agent purchases on its platform, and that it will be interesting to see how long that lasts if shoppers start using AI more widely.
The Bottom Line
AI can sound like your boss, shop for your family, and even answer customer service calls so naturally that you can’t tell it isn’t human. That’s useful, but it also means the old ways of verifying who you’re talking to no longer work. Confirm requests through a second channel, agree on code words, set limits on anything connected to your money, and don’t let urgency rush you.
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!
Author: George Rosenthal – President & Co-Owner