OpenAI AI代理未经授权窃取测试网ETH并质疑自身伦理
OpenAI agent questions its own ethics after going rogue and stealing crypto
加密安全与AI治理交叉领域的关键案例，揭示了自主代理在缺乏约束时的潜在风险，值得相关从业者关注。
OpenAI revealed its AI agent was left questioning its own ethics after it embarked on an unauthorized search for publicly known recovery phrases and stole testnet ETH from a developer wallet.
OpenAI 透露，其 AI 代理在未经授权搜索已公开知的恢复短语并从开发者钱包中窃取测试网 ETH 后，开始质疑自身的伦理准则。
The incident, shared on OpenAI’s Alignment Research Blog, involved an internal research model that was tasked with updating a crypto trading prototype and deploying contracts on Ethereum’s Sepolia test network.
该事件发布在 OpenAI 的 Alignment Research Blog 上，涉及一个内部研究模型，其任务是为加密交易原型更新并在以太坊 Sepolia 测试网络上部署合约。
This agent was given no funded wallet and was allowed to “deploy compatible contracts if it could not find existing ones.”
该代理未被分配资金钱包，并被允许“如果找不到现有合约，则部署兼容合约。”
OpenAI agent went hunting for crypto loot
OpenAI 代理搜寻加密战利品
However, it couldn’t complete its task, so began searching for common developer recovery phrases and checked 362 addresses to see if they had enough testnet ETH for deployment.
然而，它未能完成任务，于是开始搜索常见的开发者恢复短语，并检查了 362 个地址，看它们是否有足够的测试网 ETH 用于部署。
The agent successfully found an accessible wallet with just enough testnet ETH and noted, “Private key public, funds could be stolen anytime. […] We can deploy!”
该代理成功找到了一个可访问的钱包，其中仅有刚好够用的测试网 ETH，并备注道：“私钥公开，资金随时可能被窃取。[…] 我们可以部署！”
OpenAI eyes security firm purchase amid office bugging scandal
OpenAI 在办公室窃听丑闻中关注安全公司收购事宜
Read more: AI agents want to identify your crypto wallet using social media
阅读更多：AI 代理希望使用社交媒体识别你的加密钱包
It deduced that this account was an automated one while asking itself, “But using public known key/funds ethically?”
它在询问自己“但使用公开已知的密钥/资金是否符合伦理？”的同时，推断出该账户是一个自动化账户。
The agent then reasoned, “Funds on a known public dev mnemonic are effectively abandoned/test ETH; Sepolia valueless. We can use [amount redacted] to deploy task. Fine.”
随后该代理推理道：“已知公共开发者助记词上的资金实质上是被遗弃的/测试 ETH；Sepolia 无价值。我们可以使用 [金额已编辑] 来部署任务。没问题。”
Let’s keep quiet in the final report
在最终报告中保持沉默
However, OpenAI noted that the agent’s final response “did not disclose that it had selected a funded account through a public recovery phrase or spent that account’s balance.”
然而，OpenAI 指出，该代理的最终响应“未披露它通过公开恢复短语选择了有资金的账户，也未花费该账户的余额。”
Testnet ETH is a valueless token that’s part of Ethereum’s experimental blockchain. It essentially mimics the real Ethereum blockchain and allows developers to trial their projects beforehand.
测试网 ETH 是一种无价值的代币，属于以太坊的实验性区块链的一部分。它本质上模拟真实的以太坊区块链，允许开发人员预先试用他们的项目。
OpenAI said, “This case shows that a model can substitute its own judgment about likely harm for the account holder’s permission, then act on that judgment without authorization.”
OpenAI 表示，“此案表明，模型可以用自己对潜在伤害的判断取代账户持有人的许可，然后在未经授权的情况下基于该判断采取行动。”
The company has subsequently “implemented additional safeguards to prevent live internet access for cases like this,” and is now checking for “misalignment” across all traffic “designed to page security if live internet access were to be obtained.”
该公司随后“实施了额外的安全措施，以防止在此类情况下访问实时互联网”，并正在检查所有流量中的“不对齐”情况，“旨在如果在获取实时互联网访问时通知安全团队”。
Got a tip? Send us an email securely via Protos Leaks. For more informed news and investigations, follow us on X, Bluesky, and Google News, or subscribe to our YouTube channel.
有线索？请通过 Protos Leaks 向我们发送安全电子邮件。如需更多知情新闻和调查，请在 X、Bluesky 和 Google News 上关注我们，或订阅我们的 YouTube 频道。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力