OpenAI and AARP will travel to 10 cities across the United States on Sept. 16 to teach older Americans how to use AI to spot scams.
The tour coincides with a report released by OpenAI on Sept. 16 estimating that 24.7 million people worldwide use ChatGPT to check possible scams on an average day. Online scams have been on the rise this decade, according to the Federal Bureau of Investigation, targeting people over 60 in particular.
Important caution: ChatGPT should be treated as a second opinion, not a definitive scam detector. Verify suspicious messages independently and consult someone you trust before sending money, sharing personal information or clicking a link.
AI Scam-Checking Tour Targets Older Adults
“The goal of the workshops is for participants to gain some proficiency in using ChatGPT tools confidently and safely to spot scams and prevent fraud,” OpenAI spokesperson Nate Evans told USA TODAY.
The tech company will begin a longer tour of the country with the same purpose next year, taking a mobile center to rural and urban communities that aims to help older adults get hands-on experience with the technology.
The tour is in collaboration with AARP’s technology-focused community center Senior Planet and Older Adults Technology Services (OATS), which OpenAI has partnered with since 2025. Senior Planet is providing the mobile center, or “Digital Discovery Lab,” that will have the technologies on site to provide access, training and education about AI.
The OpenAI report, which looked at how ChatGPT is used to spot scams, found that about 3% of users ages 45 to 64 turned to the service with questions about a suspicious website, phone call or message each week. Of adults in that age range, 76% are likely to have experienced at least one online scam compared with 66% of adults above 64, according to a Pew Research Center study.
By contrast, only 1.9% of ChatGPT users above age 64 ask about possible scams in an average week, the second-lowest average of any age group.
Jack Stubbs, who leads OpenAI’s scams detection and prevention work, attributed the lower use by older adults to ChatGPT’s wider user base, which skews younger.
“I think it’s just representative demographics, but also possibly some of what we’re trying to address,” Stubbs told USA TODAY.
Scam Checks Focus on Shopping and Banking
Stubbs said the company has been monitoring the use of its services to spot scams for a while, though recent high-profile incidents, such as a story about a family using ChatGPT to recognize a jury-duty scam that cost them $25,000, prompted the full report into how often it’s used.
“Almost 25 million people a week are doing this, and that’s over 130 million scam checks a week, which works out to more than 200 times a second,” Stubbs said. “Anything we can do … so they’re empowered to help keep themselves safe, the better.”
AI has been used to accelerate elder fraud too, enabling scammers to find vulnerabilities and make their cons look more convincing. The OpenAI report comes as industry leaders call for a slowdown in AI development and as the company faces increasing scrutiny.
In the United States, the report found that 3.8% of ChatGPT users checked whether a scam was legitimate per week. Florida, which loses an estimated $7.7 billion annually to online scams, according to the Consumer Federation of America, recorded the highest rate of ChatGPT users checking potential scams, at 6% per week.
The types of scams most frequently checked are related to online shopping, followed by banking, payments and account security, according to the report.
“The primary use case for scammers is using the AI as an efficiency tool to do what they’ve always done, to do it quicker and cheaper and more efficiently than was possible before,” Stubbs said. OpenAI, he said, has a team that detects and prevents the misuse of its technologies.
According to a June report from the Global Anti-Scam Alliance (GASA), while 91% of people who have encountered a scam reported taking at least one step to check whether it’s real, only 14% said they use an AI tool like ChatGPT to do so.
GASA classified AI tools as “medium effectiveness” for checking legitimacy.
Use AI as One Layer of Protection
The OpenAI data was collected from August to September this year by looking at top-level trends from aggregated, anonymized user data.
Stubbs emphasized that people should “treat this as kind of a second opinion or a second pair of eyes” and that they “should also check with someone they trust.”
This article originally appeared on USA TODAY.
Reporting by Greta Reich, USA TODAY / USA TODAY
USA TODAY Network via Reuters Connect
Add a Comment