Nvidia's scandal shows the risk of relying on one leader
Nvidia is in a major PR scandal tied to CEO Jensen Huang, and it shows what can go wrong when a company leans on one leader.
Read more: Nvidia's scandal shows the risk of relying on one leaderCircle, the company behind USDC, says AI agents could pay for services within rules you set. The open question is who covers mistakes and refunds.

Circle, the company behind the USDC digital dollar, wants AI software to be able to pay for things on your behalf, as long as a person or business sets the rules first. In a November 4 explainer, it calls the idea agentic commerce. It is a vision for a developing market, not a launched product and not proof that such payments are already in wide use.
An AI agent is software that carries out tasks instead of only answering questions. Under Circle's idea, an agent could compare prices, get a computing service, pay the provider and record the purchase without a person clicking through each step. That differs from a chatbot that only suggests products, and from a subscription that renews on its own. USDC is a stablecoin, a digital token linked to the US dollar.
The hard part is not the payment. It is permission. A business would need to say which agent may spend, how much, on which kinds of suppliers and for how long. It would also need records of what triggered each payment and a way to cancel the agent's authority. Without those safeguards, a small software error or a malicious instruction could turn into a real financial loss.
Payments sent over a blockchain, the shared digital record that tokens like USDC run on, can be hard to reverse. Card payments often come with dispute rights, fraud monitoring and customer support. Those protections do not automatically come with a blockchain transfer. A workable service would need clear handling of refunds, wrong items, failed delivery and disputes. It would also have to cover sanctions checks, identity rules and local limits.
Merchants also have to accept the method, and users have to trust the controls. The most believable early uses are likely to be narrow and easy to verify, not agents managing unlimited funds.
If AI tools start spending money for you, the limits you set and your ability to get money back will matter more than how fast the payment goes through. The real test is whether automated payments get easier for ordinary businesses without making financial control harder.
Before any AI tool spends for you, check what limits you can set, how you can cancel its access and how you get a refund.
Nvidia is in a major PR scandal tied to CEO Jensen Huang, and it shows what can go wrong when a company leans on one leader.
Read more: Nvidia's scandal shows the risk of relying on one leaderStifel, a financial firm, is adding Anthropic's AI to its advisers' toolkit. Its CEO says a person still makes the calls about your money.
Read more: Stifel brings Anthropic's AI to its financial advisersDaylight, a New York security startup, raised $80 million to use AI agents to guard company systems, a sign of growing demand for outsourced protection.
Read more: Daylight, a New York AI security startup, raises $80 million