A new report just revealed exactly how much personal information your favorite artificial intelligence chatbots are pulling from your phone. Researchers at Surfshark looked at the ten most popular chatbot applications on the App Store to see what they grab behind the scenes.
The results show that not all chatbots treat your privacy the same way, and some pull a massive amount of personal details just to function.
How Meta and Google handle your information
Unsurprisingly, Meta AI grabbed the worst score on the privacy list. The app collects 33 out of 35 possible data types, which covers nearly 95 percent of everything it can track. It is the only app in the study that records financial information. Google Gemini ranked right behind it by collecting 23 different data types.
Gemini pulls things like precise location, search history, and your contact list. Meta also gathers sensitive data like political opinions and biometric details, and Google does the exact same thing with its own chatbot.
The safest way to chat on an iPhone
Across the board, the average chatbot application collects 14 different types of data. About 70 percent of them will track your location while you use the service. Upgrading to a paid subscription does not stop a developer from taking your information, either. Right now, the most secure way to ask a chatbot questions on an iPhone is through Siri.
Apple built a specific contract with ChatGPT that anonymizes your requests and blocks the service from using your inputs to train its model.
