Meta lanza modo “incógnito” en WhatsApp ante inquietudes de privacidad en chats con IA
🔒 Meta launched a new "incognito mode" on WhatsApp to allow users to have private, temporary conversations with their AI chatbot without data being saved or accessible by Meta.
⚠️ This feature addresses growing privacy concerns regarding how users share sensitive personal, financial, health, and professional information in AI chats.
🔐 Messages processed under incognito mode run in a secure environment that not even Meta employees can access.
🗑️ Unlike standard chats, these temporary conversations are not stored by default and will disappear once the user exits the session.
🤖 Generative AI systems face privacy scrutiny because large language models are trained on vast datasets, sometimes including personal user data shared during interactions.
🆚 Competitors like Google Gemini and ChatGPT already offer similar privacy controls, such as disabling chat history or opting out of data usage for model training.
🗣️ Will Cathcart, head of WhatsApp at Meta, emphasized that users increasingly ask significant questions about their lives to AI systems and may not want to share underlying information with the operating companies.
🛡️ The incognito mode includes safety features designed to prevent the chatbot from answering questions on harmful topics.
🤖 The system is programmed to guide users toward useful information when possible, refuse to answer harmful queries, and eventually stop interacting with users who violate safety guidelines over time.
📝 Users in this mode can only send text-based questions and receive text responses; image uploads or generation are not permitted.
🔞 An age confirmation requirement is enforced because Meta does not allow users under the age of 13 on its platforms.
🌐 This story was translated from English with assistance from a generative AI tool, highlighting the very technology that prompted these privacy measures.
- incógnito
- ,
- ,
- Meta faces growing competitive pressure as rival chatbot manufacturers like Google (Gemini) and OpenAI (ChatGPT) already offer privacy controls such as disabling chat history or excluding data from model training, making Meta's new feature potentially just a reactive move.
- The implementation of the 'incógnito' mode highlights deep user anxieties regarding privacy, suggesting that users may be hesitant to share sensitive financial, health, or personal data even without the feature, which could limit engagement with Meta AI.
- Will Cathcart's admission that the system will eventually stop interacting with users entirely indicates a significant limitation in the utility and scalability of Meta's AI assistant compared to competitors who may offer more robust and continuous interactions.
- Restrictions such as limiting input and output to text only, while prohibiting image uploads or generation, significantly reduce the capability and appeal of WhatsApp's AI compared to multimodal competitors.