On October 16, Charles Schwab CEO Rick Wurster confirmed on Bloomberg TV that the firm is blocking unauthorized AI agents, such as Meta Muse, from executing transactions for clients. A weekly report shows Schwab blocking dozens of attempts by major AI agents to access and manipulate client funds. The firm separates informational access from transactional authority: its in-house assistant 'Charley' cannot trade, while external partner language models get read-only access through MCP servers that bar moving money. JPMorgan CEO Jamie Dimon said his firm also blocks outside agents from accessing customer-account information. The article notes there is no federal statute on agent liability, so financial firms are setting their own read-only standards and shifting risk management back onto themselves.
Summary written by Upside AI from the original article. Read the source for full details.
Read the original on Yahoo Finance ↗
Keywords
Charles SchwabJPMorganartificial intelligenceregulation
Open a keyword to see other news on the same topic.
Related tickers and likely impact
JPM
• NeutralJPMorgan Chase
Relevance: low
| Market | US stocks |
|---|---|
| Sector | Financials |
| Industry | Diversified Banks |
| Market cap | $891.9B |
| Close on 10/16/2026 | 339.67 |
| Change that session | +1.23% |
Impact is scored by AI from the headline and may be wrong. It is not investment advice.
Related news
- Jamie Dimon Warns AI Agents That Make Payments on Your Behalf ‘May Not’ Act Properly: AI Revolution Like a ‘Tsunami’ That Must Be Managed
JPMorgan Chase CEO Jamie Dimon warned that AI agents allowed to make payments for customers would need their bank codes, effectively placing them “in your bank” with no assurance they would act in customers' interests. He cited Meta Platforms' consumer agent Muse, launched in September to handle emails, travel bookings and calendars, and said such agents could act in “treacherous” ways for the bank; on Tuesday's third-quarter earnings call he warned of possibly thousands of rogue agents. JPMorgan is still rolling out its own AI, with CFO Jeremy Barnum saying the Smart Cash tool is in early pilot; Dimon likened AI to the Industrial Revolution but moving far faster, “more like a tsunami,” that must be managed. He said AI tools are surfacing hundreds of significant vulnerabilities in open-source code, leaving industries “completely unprotected” against state actors, terrorists and criminals. Dimon also predicted AI could help treat diseases such as cancer, cut road deaths by as much as 80%, and let people work as little as 3.5 days a week within 30 years.
- Always on, always listening: New AI devices are pushing legal limits and could put users in jeopardy
Chris Thomas, a Pittsburgh business owner, wears Meta glasses, a Fieldy pendant and an Amazon Bee wristband to record and organize his daily life. A bystander at a bar objected because he could not be sure the devices were off, so Thomas now takes them off completely. More always-listening AI devices are emerging, including Amazon's Bee Frontier and Bee Eclipse and Apple's Siri Recap on Apple Watch Series 12, raising concerns about covert recording without bystanders' consent. Unlike earlier smart speakers, some of these products interpret speech without retaining the underlying audio, leaving unresolved whether this could count as unlawful interception. Attorney Colin Zick says broadly worded state laws such as Massachusetts' could apply, and in some states the wearer may face greater legal risk than the manufacturer.
- How Nvidia Could Make $150 Billion Look Small
Nvidia's board announced a $150 billion share buyback, described as the largest in the company's history, with about $250 billion in buyback capacity available now. The hosts debate whether it is enough: Rachel Warren notes it is only about 2% of market cap and reflects strong cash generation from AI infrastructure, while Tyler Crowe argues it should be at least triple, around 5% of market cap, citing $134 billion in free cash flow. The same day, Nvidia unveiled AI safety initiatives, including OpenShell, an open-source framework designed to support rival platforms from ARM and Intel, and Sentry, a hardware watchdog that can quarantine rogue software within milliseconds. Rachel Warren sees these moves as an attempt to set an industry-wide safety standard before regulators step in, while CEO Jensen Huang has said he does not want to wait for government regulation.
- Anthropic AI Safety Bill Reaches Beyond Big Tech , Sacks Warns
David Sacks warns that an Anthropic-drafted AI safety bill circulating on Capitol Hill classifies models trained above 10²⁵ floating-point operations as frontier systems, a threshold he says is written broadly enough to capture AI startups far below that scale. He argues the bill's ideas only work if tightly confined to companies genuinely operating at frontier scale. The distinction directly affects how smaller AI companies get funded and built.
- Hiltzik : Of course the government should regulate AI , but not the way the AI industry prefers
The author argues that AI companies publicly welcome government regulation, but only when it is regulation they have designed themselves. OpenAI published a “blueprint for democratic governance” in June, and Anthropic CEO Dario Amodei called for “more serious and binding regulation,” while OpenAI CEO Sam Altman has advocated a “lighter touch regulatory stance.” The author notes that regulation usually arrives after harms appear, citing the 1929 stock market crash that preceded federal securities rules, and that international oversight is hard to achieve amid geopolitical competition. He also warns that industries typically seek a seat at the table to dilute rules, pointing to Sam Bankman-Fried's push to make the CFTC the model regulator for digital assets.
- Brazil Court Votes 7 to 0 for Meta Order
Brazil's Superior Electoral Court voted 7-0 on Friday 16 October to keep in force the order issued by Minister Marco Aurélio Bellizze against Meta until the 25 October runoff. The order bars Meta from reducing reach, dropping content from recommendations or restricting search for either Lula's or Flávio Bolsonaro's political content. Lula's coalition filed the complaint on 9 October, saying Meta failed to deliver its posts and that its Meta AI assistant refused images of Lula while producing material for Flávio Bolsonaro. Meta says it takes no side, calls favouritism claims false, and attributes the problems to technical incidents and an AI safeguard gap that has since been fixed. Non-compliance carries a fine of R$500,000 (about US$98,000) a day, and Meta must restore Lula's undelivered paid boosts by Wednesday 21 October.
Related markets
Latest news
- Is Nvidia Stock Really Trading for Just 8 Times 2030 Earnings?
Nvidia is the market's most valuable company, with a market cap above $5 trillion. It trades at 27 times adjusted trailing earnings, but only about 8 times analysts' estimate for fiscal 2030 earnings, a figure the author stresses is a moving target. Revenue growth accelerated from 56% in Q2 FY 2026 to 106% in Q2 FY 2027, though guidance points to a slight slowdown in the current quarter. The stock trades at 13 times next year's earnings, and the board recently authorized a record $150 billion in share repurchases. The author also acknowledges risks such as slower growth or contracting net margins.
- US Orders Civilian Agencies to Rip DeepSeek Out of Government Systems
The U.S. Department of Homeland Security on Monday ordered civilian federal agencies to remove DeepSeek products and third-party services built on the model from their IT systems, and to stop buying them. The notice warns that using DeepSeek may transfer sensitive data to China and could produce intelligence on U.S. persons and agencies. The Commerce Department, the Navy, and the states of New York, Tennessee and Virginia had already restricted DeepSeek. The FCC votes on October 29 on barring Chinese labs from testing electronics for U.S. use.
- Tesla-SpaceX merger is 'not so easy': Analyst
Speculation about a possible merger between SpaceX (SPCX) and Tesla (TSLA) has grown on Wall Street in recent weeks. RBC Capital Markets analyst Tom Narayan says it would not be simple: SpaceX would have to acquire Tesla, likely in an all-stock deal, and would need to pay a strong premium. He sets a $480 price target for Tesla, while the shares trade near $367. The longer-term focus may be Terafab, Elon Musk's chip initiative aimed at cutting production costs by up to 90% and bypassing Nvidia and TSMC. Musk currently controls 85% of SpaceX's voting rights and 20% of Tesla's.
- Is Nvidia (NVDA) Stock Priced Right?
Chip stocks fell sharply last week, and Nvidia (NVDA) dropped nearly 2% as investors worried that heavily indebted AI companies would face higher borrowing costs after bond yields jumped for CoreWeave, a cloud company Nvidia backs. CoreWeave shares had fallen more than 15% the day before. Nvidia's own numbers show no sign of that stress: it sold $96.2 billion of chips and systems in the quarter ended July, up 106% from a year earlier, mostly from data center sales. Its net profit margin was 63.66% over the past 12 months, and management expects $108.0 billion in revenue next quarter, assuming no data center sales to China. The key risk is whether customers can keep paying for chips if interest rates stay high, and the planned partnerships to bring over $500 billion of outside capital into AI infrastructure have not yet become definitive agreements.
- Networking Stocks Rally as Tech Rebound Lifts AI Gear Makers: Ciena and Arista Climb 4%, but Cisco Barely Moves
Networking stocks rallied with the tech rebound as suppliers of AI data center gear drew buying: Ciena (CIEN) rose 4% to $347.48 and Arista Networks (ANET) climbed 4% to $183.32, well ahead of the iShares U.S. Technology ETF (IYW) at 1.29% and the Invesco QQQ Trust (QQQ) at 1.31%, while Cisco (CSCO) was practically unchanged at $107.06. Ciena's fiscal Q3 2026 adjusted EPS was $2.11 against a $1.73 consensus, with revenue up 37% year over year to $1.67 billion, and it raised fiscal 2026 revenue guidance to $6.42 billion. Its backlog reached $8.5 billion, but supply remains the limit, and management does not expect the market to return to balance before 2028. Two customers made up 41.7% of Q3 revenue, so Ciena's growth is concentrated. The next test is whether Ciena can hit its fiscal Q4 2026 revenue guidance of $1.75 billion under tight component supply.
- Gartner Predicts “80% of Neoclouds” Out of Business by 2030: Here’s Who Wins
Gartner forecast on October 14 that more than 80% of neoclouds will stop operating independently by 2030, and CoreWeave (CRWV) fell 25% last week. The article argues that hyperscalers funding AI capacity from their own cash flow, led by Amazon and Alphabet, are likely winners as credit tightens for indebted neoclouds. Amazon generated $140 billion of operating cash flow in fiscal 2025, more than covering $132 billion of capex, while CoreWeave's total debt reached $29.8 billion. Since their July 29 closing low, Microsoft, Alphabet and Amazon are up 15.6%, compared with 1.6% for IREN, CoreWeave and Nebius. Hyperscaler earnings in the coming weeks will need to show whether sold-out capacity is turning into results.
Get the briefing on Telegram or Zalo
Join the Upside Telegram channel or Zalo group for the Vietnam stocks and gold briefing at 8:00 and 19:00 (Vietnam time) on trading days, plus a weekly wrap on Saturday morning.