欧洲顶尖的银行监管机构警告指出，前沿人工智能（AI）模型（如Anthropic的Mythos）可能对金融体系构成系统性风险。欧洲中央银行（ECB）已向110家银行发出信函，要求他们在10月底前提交一份应对AI带来之网络安全威胁的全面行动计划。

欧洲系统风险委员会（ESRB）表示，新一代AI模型让黑客能以极低成本自动执行高度复杂的网络攻击，将系统性网络风险评级从「升高」提升至「严重」。同时，英格兰银行行长安德鲁·贝利虽认同AI带来的重大威胁并呼吁国际合作，但表示英方将采取与银行密切合作的温和作风，而非强行下达法令。

欧盟委员会随后也提交了针对先进AI模型的网络安全行动计划，预计将于明年建立评估机制以保护关键基础设施。为了让银行有充足时间应对紧迫的期限，欧洲央行已将年度IT风险问卷的提交截止日期从9月延后至明年2月。

Europe's top banking watchdogs have warned that frontier artificial intelligence models, such as Anthropic's Mythos, could pose systemic risks to the financial system. The European Central Bank has sent a letter to 110 lenders, giving them a tight deadline until the end of October to come up with a comprehensive action plan to combat these AI-related cyber risks.

The European Systemic Risk Board stated that modern AI models allow hackers to cheaply automate highly complex attacks, raising the systemic cyber risk rating from "elevated" to "severe." While acknowledging the significant threat and calling for international cooperation, Bank of England Governor Andrew Bailey noted they are taking a collaborative approach with banks rather than issuing edicts.

The European Commission also presented its action plan to address cybersecurity risks from advanced AI, intending to establish an evaluation capacity next year to protect critical infrastructure. To help banks cope with the tight timeline, the European Central Bank has postponed the deadline for its annual IT risk questionnaire from September to February. (Key numbers: 9, 2)