Google AI Expert Warns Users On Cybersecurity Risks, Urges Safeguarding Sensitive Data From Public Chatbots

A Google AI security professional has shared critical advice on how users can safeguard personal and work data while using AI chatbots. Treat AI Chatbots Like A Public Postcard To Prevent Data Leaks Speaking about the rise of AI in daily life, Harsh Varshney, who has worked on Google's privacy and Chrome AI security teams, said, "AI models use data to generate helpful responses,” reported Business Insider. “We users need to protect our private information so that harmful entities, like cybercriminals and data brokers, can’t access it." They emphasized avoiding the sharing of Social Security numbers, credit card information, home addresses, or medical records with public AI tools, which may store data for future model training. Use Enterprise AI And Regularly Delete History The engineer also highlighted the importance of using enterprise-grade AI tools for confidential work conversations. "Once, I was surprised that an enterprise Gemini chatbot was able to tell me my exact address," they said, explaining how long-term memory features can retain previously shared information. They recommended deleting chat history regularly and using temporary or "incognito" modes to minimize risks. Finally, they advised sticking to well-known AI platforms and reviewing privacy settings to prevent conversations from being used in training models, ensuring both convenience and security. See Also: Ronald Reagan ‘Didn’t Love Tariffs,’ Says Economist Paul Krugman: He Repeatedly Emphasized ‘The Virtues of Free Trade’ Generative AI Privacy Risks: Which Platforms Protect Your Data? Earlier this year, a surge in privacy concerns emerged as generative AI and large language models became widely used. A report from Incogni found Mistral AI’s Le Chat ranked safest, followed by ChatGPT and Grok, thanks to clear privacy policies and opt-out options. Conversely, Meta Platforms Inc.’s (NASDAQ:META) Meta AI, Alphabet Inc.’s (NASDAQ:GOOG) (NASDAQ:GOOGL) Gemini, and Copilot from Microsoft Corp. (NASDAQ:MSFT) were the most aggressive in data collection, often lacking transparency. Mobile apps mirrored these trends, with Le Chat, Pi AI, and ChatGPT posing the lowest risks, while Meta AI collected sensitive data like emails and locations. Users were advised to review privacy settings to safeguard personal information. Read Next: Microsoft CEO Says AI Sector Needs Multiple Winners: ‘Otherwise It’ll Be A Road To Nowhere’ Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.