Artificial intelligence company Anthropic has announced a partnership with password manager 1Password, which will allow its AI chatbot to perform various online tasks with the user’s permission.

According to the company, the cloud will securely use credentials through 1Password without directly accessing the user’s login information. In this way, passwords or one-time passcodes will neither be stored in the AI ​​model nor become part of its memory.

With this new feature, users will be able to entrust the cloud with travel booking, online shopping, managing various accounts and other tasks that require logging in.

1Password says that the cloud is given limited permissions only for the current task and this access is automatically terminated as soon as the task is completed. Moreover, after autofill, the system also checks that sensitive information has not been exposed or leaked on the web page.