美国第三大商业银行花旗集团(Citibank)本周更新了在线用户协议,明确告知使用第三方AI工具进行个人财务管理的客户需自行承担所有风险 1。新条款规定,若客户授权第三方AI代理访问账户,客户需对通过该代理进行的所有活动负责,即使发生黑客攻击或凭证泄露导致损失,银行也不承担责任 1。花旗集团在声明中表示:“我们最近更新了在线用户协议,以提供关于这些风险的更多清晰度,并重申选择授予第三方AI代理账户访问权限的客户需对该代理进行的活动负责。” 1
此举紧随其他大型银行之后,反映出传统金融机构试图通过法律条款界定新型技术带来的责任边界 1。摩根大通(JPMorgan)和美国银行(Bank of America)上月已宣布类似政策;PNC银行(美国第八大银行)今日也更改了相关条款 1。此外,Citizens银行本周向客户发送邮件,警告其对AI助手的操作负责 1。
这一趋势反映了银行对AI进入核心业务领域的警惕,视AI工具为竞争威胁,担心其帮助用户规避费用或利用低息存款账户从而侵蚀利润率 1。值得注意的是,Anthropic推出的个人金融工具被明确标注为“只读”(read-only),不能移动、管理或投资资金 1。
Citibank, the third-largest commercial bank in the United States, updated its online user agreement this week to clarify that customers who grant third-party AI agents access to their accounts are responsible for all activities conducted through those tools 1. The new terms explicitly state that the bank will not accept liability for losses resulting from hacking incidents or credential leaks involving these automated assistants 1. This update reflects a broader effort by traditional financial institutions to define legal boundaries and mitigate risks associated with AI integration in personal finance management 1.
In its statement, Citibank emphasized that the revised agreement aims to provide greater clarity on these risks and reaffirms that clients granting account access to third-party AI proxies remain accountable for the agent’s actions 1. This move follows similar policy announcements by JPMorgan Chase and Bank of America last month 1. PNC Financial Services, the eighth-largest U.S. bank, also changed its terms today to adopt a similar stance 1. Additionally, Citizens Financial Group sent emails to customers this week warning them that they are responsible for operations carried out by AI assistants 1. While Anthropic has labeled its personal financial tool as "read-only," preventing it from moving, managing, or investing funds, banks continue to view such technologies as competitive threats capable of eroding profit margins by helping users avoid fees or low-interest deposit accounts 1.
评论
还没有评论,欢迎留下第一条。