Meta wants to help users feel more comfortable engaging with its artificial intelligence chatbot on WhatsApp, so it’s incorporating a new Incognito Chat mode that will utilize private processing to ensure that any in-app conversations users have with its chatbot will remain private and protected.
Meta’s Incognito Chat on WhatsApp better aligns with WhatsApp’s privacy approach, ensuring that whatever users share with the chatbot remains private, and cannot be read by Meta, even for the assessment of accuracy.
As explained by Meta: “Chatting with AI has quickly become a critical part of how people get information and ask important questions. And many of these questions can be deeply sensitive, or include situations where people are including private financial, personal, health or work data with their questions.”
Indeed, the use of AI chatbots is clearly on the rise, with users prompting AI tools with a broad range of queries.
But in most cases, those queries are also used for model training, and refining both systematic understanding and response, in order to improve the accuracy and relevance of AI tools.
With WhatsApp’s incognito AI chat mode, Meta says that won’t happen, so users can feel free to ask whatever they like, without fear of other people knowing about their concerns or questions.
“When you start an Incognito Chat with Meta AI, you’re creating a private, temporary conversation that only you can see. Your messages are processed in a secure environment that even Meta cannot access,” Meta said. “Your conversations are not saved and by default, your messages disappear — giving you a space to think and explore ideas without anyone watching.”
Which could also be problematic, in that if nobody’s checking these AI-generated responses, accuracy could be a concern.
AI chatbots are good at providing answers based on sourced information, but they’re not perfect, and they still generate hallucinations and false sources in response to some user queries. Presumably, Meta AI is still in development in this respect, and as such, enabling users to consult the chatbot with no oversight seems like Meta may be getting ahead of itself in terms of building user trust, especially for sensitive queries.
What if people use this option to get advice on how to build a bomb, how to make drugs or how to launder money? What if Meta’s AI tool offers incorrect health advice, and no one has any insight into its responses?
While the private use case makes sense, there are expanded concerns that do require a level of oversight, and it seems reckless for Meta to be pushing ahead with this option at this stage.
In addition to incognito mode, Meta said that it’s also developing a new Side Chat option for AI chats on WhatsApp, which will enable users to access its AI chatbot for assistance within any chat, without disrupting the main conversation.
The utility case is clear, in that more people are using AI to find answers to questions. But the accuracy aspect is a concern.
