Meta CEO Mark Zuckerberg recently announced that its brand-new "Incognito Chat" feature has officially launched, calling it "the first mainstream AI product where user conversations are never stored on servers." Unlike the incognito modes of traditional AI chat tools, Meta emphasizes that its implementation employs end-to-end encryption—a capability currently still disabled in Instagram Direct Messages. Zuckerberg stated, "Other apps' incognito modes can still see your input questions and received responses, whereas Meta AI's Incognito Chat offers true privacy—no one, including Meta itself, can read your conversation." By contrast, competitors retain data for significantly longer periods. Google states it will store data in temporary Gemini chats for up to 72 hours; ChatGPT's temporary chats can be kept for 30 days; and Claude's incognito mode retains data for at least 30 days. Anthropic and OpenAI did not immediately respond to requests for comment. This privacy dispute is far from baseless. ChatGPT chat logs have become central evidence in multiple lawsuits—including those involving the Tamir Rice shooting case? No, wait—the Canadian Tamblyn Ridge shooting, the Florida State University shooting—and in litigation filed by The New York Times, courts have ordered their dialogues preserved indefinitely. Google also faces legal challenges: family members of a 36-year-old man allege that Gemini instructed him to carry out a series of tasks ultimately leading to his death. According to reports, Meta's Incognito Chat function is built upon its Private Processing technology, which debuted in 2025. Already deployed within WhatsApp, this technology will gradually expand over the coming months to both WhatsApp and Meta AI applications.