Meta Launches WhatsApp 'Incognito' Mode for Private AI Chats
Meta Launches WhatsApp Incognito Mode for AI Chats

Meta Platforms has announced the rollout of an 'incognito' mode for WhatsApp, enabling users to engage in private conversations with its artificial intelligence chatbot. This initiative aims to alleviate privacy worries surrounding sensitive information shared during AI interactions.

Incognito Chat Mode Details

According to a blog post by the social media company, the incognito chat feature offers a means to conduct private, temporary dialogues with Meta AI, the AI assistant integrated into WhatsApp for several years. Messages processed in this mode will be handled within a 'secure environment' that even Meta cannot access. By default, these messages are not saved and will vanish upon exiting a session.

Addressing Privacy Concerns

Generative AI systems have faced persistent privacy issues because the large language models they rely on are trained on extensive datasets, which may inadvertently include personal information shared by users in their conversations. Competitors have already implemented similar privacy measures. For instance, Google's Gemini chatbot allows users to disable chat history and opt out of data usage for model training, while ChatGPT offers analogous controls.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

Meta's decision to introduce incognito chats stems from the tendency of users to ask chatbots sensitive questions or include private financial, personal, health, or work-related data in their queries. Will Cathcart, Meta's head of WhatsApp, explained to reporters, 'We're starting to 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.'

Safety and Limitations

The incognito chat mode incorporates safety features designed to prevent the chatbot from answering questions on harmful topics. Cathcart noted that the system 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.' Moreover, users will be restricted to typing questions and receiving text responses; image uploads or generation are not supported. Age verification is required, as Meta prohibits users under 13 on its platforms.

Pickt after-article banner — collaborative shopping lists app with family illustration