New Yorkers should leave Social Security numbers, passwords, medical records and other sensitive information out of conversations with AI chatbots, state consumer protection officials warned Monday.
The New York Department of State’s Division of Consumer Protection issued the alert as people increasingly turn to chatbots for information, health advice, emotional support and companionship. The division warned that disclosing private details can increase the risk of identity theft and financial harm.
“AI chatbots are helpful, but sharing sensitive information can increase risk of identity theft and other financial harms for users,” Secretary of State Walter Mosley said.
The warning concerns information people voluntarily type or upload, including account credentials and documents that contain identifying details. Public chatbot services may collect prompts, retain conversations or use them to train future AI models, depending on the service and its settings.
What information should stay out of a chatbot?
The division advised users against sharing Social Security numbers, names, addresses, birth dates, financial account numbers, medical diagnoses, laboratory results, legal documents, usernames or passwords. Acting state Chief Information Officer Jenson Jacob also cautioned against entering driver’s license and passport information, PINs and credit card details.
“The most effective safeguard to secure your data is to never share Social Security numbers, driver’s licenses, passport numbers, account passwords, PINs, credit card details, health information, or other personal, sensitive information when utilizing these platforms,” Jacob said.
The alert cited Stanford research on chatbot privacy policies examining 28 policy documents from six major U.S. companies. The research identified potential exposure through data retention, model training and human review, with practices varying among services. Information retained by a company can also be vulnerable to a data breach.
Chatbot use has become widespread. A Pew Research Center report published in June found that 49% of U.S. adults used AI chatbots and 24% used them daily.
Pew surveyed 5,119 adults Feb. 17-23 through its nationally representative American Trends Panel, using online and telephone interviews. Results were weighted to reflect the adult population; the full sample’s margin of sampling error was plus or minus 1.6 percentage points.
- TeraWulf takes Lansing to court over records requests as data center ban advances
- Google clarifies school AI policy, but dispute over student searches remains
- Surveyed supply-chain leaders put delivery speed ahead of price
- Environmental group plans Penn Yan protest, pushes Torrey data center moratorium
Check privacy settings and understand private browsing
The division recommended reviewing each chatbot’s privacy and training controls. Where a service permits it, users can opt out of having future conversations saved or used for training by changing the platform’s default settings.
Private or incognito browsing is a separate issue. Those modes can reduce information stored on a person’s device during internet searches, but they do not prevent information from being transmitted to outside parties, including AI services, the division cautioned.
The Division of Consumer Protection offers voluntary mediation when a consumer has been unable to resolve a dispute directly with a business. Its Consumer Assistance Helpline, 800-697-1220, operates from 8:30 a.m. to 4:30 p.m. Monday through Friday, excluding state holidays.
Consumers can file complaints at any time through the division’s website, which also provides consumer protection resources and alerts.
FingerLakes1.com is the region’s leading all-digital news publication. The company was founded in 1998 and has been keeping residents informed for more than two decades. Have a lead? Send it to news@fingerlakes1.com.