banking

BofA Global Research warns that a sustained correction in the AI trade poses a major risk to the global economy, as data-center backlash becomes a ballot issue in 10 states ahead of the midterm vote. Meanwhile, data-center construction spending hits record levels and big tech capex nears $800 billion this year.

A practitioner asks whether AI agents can handle large legacy migrations (Java 8 and outdated Spring upgrades) at insurers or banks, weighing agent autonomy against bank-grade change-control processes and highlighting Orq AI's scoped-workflow software factory as a candidate approach.

Canada's major banks — RBC, TD, CIBC, and BMO — announced they will block personal AI agents such as Meta's Muse and OpenAI's ChatGPT from accessing clients' online bank accounts, citing safety and security concerns amid reports of unauthorized third-party AI agent activity in accounts at major U.S. banks.

Natural introduces Direct in general availability, letting AI agents programmatically choose between payment rails like ACH, Wire, and Realtime (RTP/FedNow) at runtime from a single Natural wallet, with existing controls and limits still applying.

A cybersecurity professional at a Canadian financial institution asks how banks are detecting and blocking third-party AI agents (such as Meta's Muse) from accessing customer accounts, discussing WAF rules, bot management, robots.txt/ai.txt, OAuth DPoP token binding, and application-level controls, and soliciting real-world implementation experiences from the community.

Australia's Commonwealth Bank reversed its decision to replace 45 call centre workers with an AI voice bot after call volumes rose, illustrating Gartner's finding that only 14% of customer service issues are fully resolved through self-service.

Bank of America emailed customers warning them not to share banking logins with AI assistants or agents, highlighting growing security risks as personal AI agents gain popularity and banks struggle with unauthorized agent access to customer accounts.

纽约时报报道，若 Anthropic 的 IPO 仿效 SpaceX 模式，高盛与摩根士丹利各可赚取约 1 亿美元承销费；高盛取得主承销商地位，而摩根士丹利此前还参与为 Anthropic 芯片销售提供融资的博通 630 亿美元交易。

Simon Taylor maps AI agent protocols across entry, identity, tools, and intent layers, arguing that banks and merchants need verified identities, limited permissions, and audit trails to safely authorize agent actions, while cautioning that access fees could slow adoption.

pwn.watch 实时安全事件汇总：美国主要银行发现客户账户中存在未经授权的第三方 AI 代理，GitLab 披露 CVSS 10.0 未认证 RCE 漏洞（CVE-2026-110264），以及多个未修补的 Linux 内核高危漏洞细节被意外公开至公共邮件列表。

Barclays announced it is cutting around 5,000 jobs, about 5% of its global workforce, as it increases its use of AI in business operations, with cuts expected to be completed by year-end. Unions criticized the move as using AI as an excuse to discard workers, while analysts noted broader implications for the banking industry.

A Google Cloud outage disrupted banking services and numerous other operations across Europe, highlighting reliance on cloud infrastructure for critical financial systems.

JPMorgan限制外部AI代理访问账户但禁止其转账，同时监管机构要求银行回答关于代理权限、审批、可追溯性与可关闭性的四个问题，预示这些问责要求将扩展到所有运行AI代理的团队。

JPMorgan, Citi, Bank of America, and Schwab moved to block external AI agents from accessing customer accounts or moving money, keeping only read-only access via official connectors. The author argues this enforces at the institutional level the approval-gate permission architecture that agent frameworks should build in from the start.

The Consumer Financial Protection Bureau has opened a debanking investigation into nine large national banks over account closures related to crypto and political views.

巴西最大银行 Itaú 与 OpenAI 合作，打造覆盖约 7000 万客户的生成式 AI 平台 ia.i，以及用于模型治理与编排的内部平台 IARA，将 AI 深度融入零售金融、法律文档处理等业务场景。

Show HN 介绍 Stoke Money，一款围绕现金流预测而非传统预算构建的个人理财应用，自动检测经常性交易并逐日推演账户余额，Pro 版支持通过 MCP 连接 Claude/ChatGPT 等 AI 助手查询与管理数据。

JPMorgan、摩根士丹利、SMBC及MUFG等银行正寻求将此前承销的AI数据中心贷款风险转售给外部投资者，显示出贷款机构对AI基建债务的敞口已达内部限额，Oracle信用违约互换升至2009年以来高位。

中国开发者在韩国银行遭黑客攻击事件后，将其 ARTEX AI 智能体转为闭源。

CrowdStrike 报告显示，上周攻击多家韩国大银行的网络攻击可能由单人完成，攻击者组合使用了开源 AI 渗透工具 ARTEX 以及 DeepSeek v4.1-Flash、GLM-5.3、Grok 4.6 和 Claude Code 等多个 AI 模型完成整个入侵过程。