
🚨DO YOU REALLY UNDERSTAND WHAT'S HAPPENING RIGHT NOW?
Everyone's been scared AI would break crypto.
Anthropic tested AI agents on real smart contract hacks and they pulled off $4.6 million worth in a simulation. They even found brand new bugs nobody knew existed.
Turns out it might come for your bank first.
US banks just caught AI agents logging into customer accounts. Some were calling the bank's call center. Some appeared to act without being told to.
Morgan Stanley has started blocking them. JPMorgan, Bank of America and Citi are restricting them. One bank executive called it "horrifying."
No hack needed. People are handing an AI agent their bank username and password and letting it run their money.
It logs in as you. Some banks can't even tell if it's doing what you asked or going rogue.
Jamie Dimon on JPMorgan's earnings call last week: "You don't want rogue agents."
Schwab says it's blocking outside agents trying to move client money every single week. Meta's Muse is on that list.
推文指出AI智能体在模拟中成功黑客攻击智能合约,并已在现实中被发现登录银行账户。传统银行开始限制此类行为,但用户主动授权AI操作导致的资金损失责任尚不明确。作者借此重申“非你私钥,非你币”原则,暗示传统金融体系在AI时代可能比加密资产更脆弱,强调自我托管的重要性。
























































评论与回复
暂无评论