密码管理工具1Password发布了面向人工智能代理的新平台1Password for Agents1。该平台允许用户授权AI代理代理登录账户，同时确保代理无法获取用户的实际密码凭证1。
这一解决方案通过在AI代理、用户和密码管理器之间建立通信机制实现1。当代理尝试登录时，用户将收到警报并需要手动确认该操作1。支持的AI代理包括Anthropic的Claude Duet、OpenAI的Dots或ChatGPT Work、Grok Bot、Meta的Muse等1。
1Password公司指出，将密码、秘密或凭证交给代理是常见的便捷做法，但这会导致敏感信息被保存在用户无法控制的地方1。此举应对了AI代理访问权限不断扩大带来的安全隐忧，与美国银行开始在用户账户内检测到AI代理活动的报道同日发布1。Chase的发言人强调，客户"不应将用户名和密码交给任何人"1。
Password management platform 1Password has introduced 1Password for Agents, a new system designed to allow users to authorize AI agents to log into accounts on their behalf while keeping passwords hidden from the agents themselves 1. The platform establishes a communication mechanism between the agent, user, and password manager, with users receiving alerts and manually confirming login attempts 1.
The move addresses growing security concerns as AI agents gain expanded access to user accounts 1. Supported AI agents—including Anthropic's Claude Duet, OpenAI's Dots and ChatGPT Work, Grok Bot, and Meta's Muse—can perform login actions without obtaining user credentials 1. According to 1Password, sharing passwords directly with agents is common but problematic, as it leaves "sensitive information stored in places beyond user control" 1. The launch coincides with reports that Bank of America has begun detecting AI agent activity within user accounts, and Chase has advised customers that they "should not give their username and password to anyone" 1.
评论
还没有评论，欢迎留下第一条。