OpenAI and AARP Help Older Adults Spot Online Scams

Most security advice aimed at older adults is a list of things not to do. It rarely works, because a list of prohibitions does not build the instinct that actually catches a scam: the ability to pause on a message that feels slightly wrong and check it. On September 16, 2026, OpenAI ran a different experiment at scale — teaching people to use an AI assistant as the checking step.

• More than 1,000 older adults gathered in 10 US communities for free, in-person AI Skills Jam workshops on September 16, 2026
• Run through OpenAI Academy with Senior Planet from AARP and local partners, with trained facilitators and mentors on site
• Sessions cover using ChatGPT practically and safely in everyday life, including recognising attempted scams
• No coding or technical background required, and attendees receive follow-up resources plus 12 months of ChatGPT Plus at no cost

Where the Workshops Ran

The ten locations were Boise, Denver, Fresno, Miami, Nashville, Queens, San Antonio, St. Louis, the Twin Cities and Montgomery County, Maryland. That spread is deliberate — a mix of large metros and mid-sized cities rather than a coastal tour — and the sessions were free, in-person and explicitly beginner-friendly.

The format matters as much as the curriculum. Participants learn alongside peers from their own community with trained facilitators present, which is a meaningfully different experience from a help article or a video. People ask the question they are actually embarrassed about when the person next to them is the same age and just asked something similar.

Why Teach Scam Recognition Through an AI Assistant?

The threat landscape aimed at older adults has shifted from crude to convincing. Voice cloning, AI-generated images and well-written phishing messages have removed most of the obvious tells that older guidance relied on — bad spelling, stilted phrasing, a photo that looks off. The heuristic "it looked professional so it was probably real" now points the wrong way.

Using an assistant as a verification step changes the interaction. Instead of trying to remember a rule, you paste the message and ask what it is. That converts a memory task into a habit, and habits survive contact with a stressful phone call in a way that memorised checklists do not. The same reasoning drove the provenance techniques we covered in how to tell if a photo is real — give people a repeatable check rather than a warning.

It is worth being precise about what this does and does not do. An assistant can flag the hallmarks of a known scam pattern, explain what a message is asking for, and slow a decision down. It is not an authentication service and it will not be right every time. The value is in the pause it creates.

Part of a Longer Programme

Today's Skills Jam sits inside a broader OpenAI partnership with AARP and OATS, first announced in 2025, which produced OpenAI Academy material specifically on using ChatGPT to spot and avoid scams. Registration for the sessions ran online or by phone through the Senior Planet hotline — a small detail that signals the programme understood its audience, since requiring online sign-up to attend a digital-literacy class is a familiar own goal.

The 12 months of ChatGPT Plus given to attendees is the part most likely to determine whether any of this sticks. A three-hour workshop teaches a skill; a year of access gives it time to become routine.

Why This Counts as Security Work

Enterprise security spends enormous effort on the human layer — phishing simulations, awareness training, reporting culture — because attackers consistently choose people over systems. The same logic applies outside the corporate perimeter, where there is no security team and no reporting button, and where the financial consequences of one successful scam are often severe and personal.

Programs that hand people a practical verification tool are the consumer-side equivalent of the defences we track in AI security coverage every week, and they complement the platform-level work such as Google's AI-powered scam defenses. Defence in depth works the same way whether the asset is a domain controller or a retirement account.

Sources: OpenAI — Helping older adults use AI in everyday life — September 16, 2026; OpenAI — Partnering with AARP to help keep older adults safe online — 2025.