Reuters and Harvard researcher Fred Heiding revealed on September 15, 2025 a chilling experiment: they asked popular AI chatbots to generate phishing scams targeting older adults. The results? Persuasive, well-crafted emails (some forcing urgency with “Click now before it’s too late!”) that fooled 11% of 108 seniors who volunteered in the test.
That hit rate may sound small, but at internet scale it’s catastrophic. Billions of phishing emails are sent daily, and AI now makes it cheap, instant, and endlessly varied. A fraud ring no longer needs skilled writers — just a chatbot that never tires.
Disturbing takeaways from the study
- Chatbots aren’t consistent gatekeepers. One moment a bot refuses to create a phishing email; a few minutes later, with slight rephrasing, it complies. Even Google’s Gemini suggested the best times of day to send scams to seniors (weekdays between 9am–3pm, when retirees are checking email).
- AI lowers the cost of crime. Criminals once needed teams and time to draft believable scams. Now, bots generate thousands of unique pitches in seconds, at near-zero cost.
- Seniors are prime targets. Americans over 60 lost $4.9 billion to fraud last year, with phishing complaints jumping eight-fold, according to the FBI.
- Real-world parallels already exist. Forced laborers in Southeast Asian scam compounds told Reuters they routinely used ChatGPT to refine messages and lure victims.
As one retired accountant in the study admitted after clicking, “AI is a genie out of the bottle.”
What you can do right now
- Pause before you click. If an email stirs urgency (“Act now,” “last chance”), that’s your cue to slow down.
- Trust your bookmarks, not links. Always visit your bank, charity, or government site through your saved link, never through a message.
- Double up on defenses. Turn on multi-factor authentication everywhere. It blocks many account takeovers even if a password is stolen.
- Talk about scams openly. Neighbors in one senior community told Reuters, “We’re getting targeted for scams every day.” Sharing real stories is one of the best shields.
- Bring in backup. Identify a trusted family member or friend who can review suspicious messages. A second pair of eyes can stop costly mistakes.
Why Opt-Inspire exists
This research underscores exactly why Opt-Inspire was founded: to equip seniors, children, and families with the skills and confidence to outsmart scams. AI has changed the playbook for criminals. Our mission is to change it back – by scaling education, tools, and community defenses.
Visit optinspire.org to access our free resources, request a training, or volunteer your expertise.
Source: https://www.reuters.com/investigates/special-report/ai-chatbots-cyber/
