Meta Platforms on Wednesday said it is rolling out an “incognito” mode for WhatsApp users who want private, temporary conversations with its Meta AI chatbot, addressing concerns about how sensitive information shared in AI chats is handled. The company said the option is intended to ease worries that personal details provided during conversations could be stored or seen by the company operating the chatbot.
In a blog post, Meta said incognito chat mode provides a way to have private, temporary conversations with Meta AI within WhatsApp, where the assistant has been available for a few years. Meta described the mode as processing messages in a “secure environment” that even Meta can’t access, alongside default settings that will not save messages.
Meta also said incognito chats will disappear when users exit a session. The company said its approach is designed to reduce how much users feel they need to reveal the information behind their questions to the companies that run AI systems.
Meta’s announcement comes amid broader privacy concerns about generative AI systems, the company said, because the large language models that power these assistants are trained on vast data sets that can include personal information users provide in their own conversations. Meta said it is rolling out incognito mode because people often ask chatbots sensitive questions or include private financial, personal, health or work information in prompts.
Will Cathcart, head of WhatsApp, told reporters: “We’re starting ask a lot of meaningful questions about our lives with AI systems, and it doesn’t always feel like you should have to share the information behind those questions with the companies that run those AI systems.” Cathcart also discussed safety features for incognito chats, saying they are intended to limit harmful responses.
Meta said the chatbot will “steer the user towards helpful information if it can and then refuse (to answer) and eventually even just stop interacting with the user completely,” according to Cathcart. The company said users in incognito chat mode can type questions and get text responses, but they won’t be able to upload or generate images.
Meta also said users will have to confirm their age because it does not allow users under 13 on its platforms. The company pointed to existing privacy controls from rival chatbot makers, including Google’s Gemini option to disable chat history and opt out of allowing data to be used in training, and controls in ChatGPT with similar effects.