Meta AI collected 32 of 35 different data points from its users, including contact details, financial information, user content, usage details, diagnostics, and location, making it the most aggressive among a group of 10 widely used artificial intelligence chatbots in obtaining user data, ZDNET reports.
Google Gemini was the second most user data-hungry AI chatbot, followed by Poe, Claude AI, and Microsoft Copilot, with the latter being the only other chatbot aside from Meta AI that gathered user identity-linked information to show third-party advertisements, according to a Surfshark analysis. Moreover, both Poe, Copilot, and Jasper were discovered to have obtained data leveraged for user tracking, with the first two also having peddled user's device IDs to data brokers. Chinese AI platform DeepSeek, which has raised data privacy concerns due to its China-based servers, only ranked sixth but researchers have cautioned about the prolonged data storage policy of the firm. "Don't let your guard down, as chats stored on servers are always at risk of being breached," said Surfshark.
Google Gemini was the second most user data-hungry AI chatbot, followed by Poe, Claude AI, and Microsoft Copilot, with the latter being the only other chatbot aside from Meta AI that gathered user identity-linked information to show third-party advertisements, according to a Surfshark analysis. Moreover, both Poe, Copilot, and Jasper were discovered to have obtained data leveraged for user tracking, with the first two also having peddled user's device IDs to data brokers. Chinese AI platform DeepSeek, which has raised data privacy concerns due to its China-based servers, only ranked sixth but researchers have cautioned about the prolonged data storage policy of the firm. "Don't let your guard down, as chats stored on servers are always at risk of being breached," said Surfshark.
