By realmnova 
At the 2026 WWCC Development Conference, Apple launched a new feature called “Apple Inteligence”, whereby AI can automatically change the password that may be attacked by hackers.
This function is based on the safety precautions previously introduced by the apple to identify weak or leaked passwords. According to Apple, Apple Intelligence was able to access and log on the Apple ‘ s AIT-based website, thereby enabling the Apple smart system to “act on behalf of the user ” . The system will then keep the new password in the password application so that users do not worry about changing the password.
Previously, in order to repair the leaked passwords, users needed to access each site manually and change the passwords on a case-by-case basis in order to act accordingly. This new feature eliminates this cumbersome step by automatically completing the entire process backstage.
Apple describes the system as an intelligent system, using Apple Intelligence and Safari to view the website safely, log in the account and upgrade the account to a strong password, and the user can do so only a little bit later. This feature will be displayed as a “real-time activity” when activated.
Although Google has built-in tools such as Google Password Manager to detect weak or leaked passwords, Gemini is currently unable to change them for the user, making the tool quite novel in the application of proxy AI.
This function also feels like a different type in the proxy AI area, as the terms “cipher” and “AI agent” are usually placed in the same context, usually the trigger for all security alerts. For example, OpenClaw has more than once been described as a security nightmare.
It is worth mentioning that the recent hole in Meta AI chat robots has led to the theft of more than 20,000 Instagram accounts and the theft of the original master account by the outlaws by seducing AI to reset the password to the mailbox provided by the hackers.