What's happened
Privacy experts warn that AI tools like Google Gemini and OpenAI ChatGPT are surfacing personal contact information without consent, prompting concerns about privacy breaches and new fraud vectors. Incidents cited include a Reddit user’s phone number appearing as a business placeholder and scammers planting fake numbers linked to AI chatbots. Companies say safeguards are in place, but incidents persist as data brokerage histories meet generative AI.
What's behind the headline?
What this means for users
- AI systems are now accessing and regurgitating contact data that exists online, sometimes as public records, sometimes through data brokers.
- Breaches can happen even when queries are benign, because providers feed models large, varied datasets and may rely on third-party data sources.
- The risk extends to customer-service interactions, where fake numbers can misdirect users to scams rather than legitimate providers.
Who is driving this
- Privacy firms and researchers are highlighting systemic data-trading practices that feed AI training.
- Tech companies defend safeguards but face ongoing scrutiny as incident reports rise.
The likely near-term outcome
- Expect increased demand for privacy controls, opt-outs from data used to train models, and clearer disclosure about data sources. Regulators may intensify scrutiny of data brokerage and AI training pipelines.
How we got here
Experts say years of data brokerage have fed into AI training data, creating unprecedented risk when systems surface or fabricate real contact details. Reports show researchers exposing personal numbers and addresses via AI queries, with industry players acknowledging ongoing privacy challenges as AI assistants become more integrated into everyday services.
Our analysis
New York Post reports on AI doxxing and data brokerage; The Independent and MIT Technology Review provide context on how AI systems surface personal data; AV and security researchers are cited by MIT Technology Review; Virgin Media O2 case noted by press outlets; Google comments on safeguards.
Go deeper
- Will these incidents prompt stronger privacy rules for AI training?
- How should individuals protect personal data when using AI chat tools?
- What safeguards are tech providers planning next to curb doxxing?
More on these topics
-
Meta - Social media company
Facebook, Inc. is an American social media conglomerate corporation based in Menlo Park, California. It was founded by Mark Zuckerberg, along with his fellow roommates and students at Harvard College, who were Eduardo Saverin, Andrew McCollum, Dustin Mosk