Elon Musk made a bold statement regarding the new AI agent Grok. According to him, if the bot, while managing a user's bank account, causes a loss of funds, the company xAI will fully compensate for the losses. However, as documents show, the company's actual liability is limited to $100, which calls into question these loud guarantees.
Experiment with a bank account: a test of strength
One investor, known by the nickname Teslaconomics, decided to test Grok's capabilities in practice. He asked on the social network X whether anyone had connected the bot to their bank account. Teslaconomics noted that the AI agent can track expenses, pay bills, and detect suspicious transactions. At the same time, he wondered whether he now needed a personal banker. However, his colleague expressed concern that giving AI access to finances could lead to unpredictable consequences.
Musk was not fazed by such doubts. He is confident in his product and even promised to refund money in case of a bot error. The beta version of the service appeared on August 11. At xAI, they explained that each agent operates around the clock on its own cloud server. It visits websites as a regular user and continues working while the owner sleeps. This experiment fits into Musk's global financial strategy: in June, the X Money service with direct transfers between users already went live.
Liability limitation: $100 — and no more
Musk's promise sounds like insurance, but the terms of the xAI user agreement say otherwise. It states that the results and actions of agents are provided "as is." The maximum claim amount cannot exceed the amount of paid fees or $100 — whichever is greater. Access to the bot costs $30 per month under the SuperGrok plan, that is, $360 per year. But this is a trifle compared to the amount that could be lost from an account.
Musk's response on X does not change these terms. Until xAI formalizes the guarantee in writing, any compensation remains at Musk's discretion. The risks are also compounded by US banking rules. The country has Regulation E — a federal standard for electronic transfers that protects customers from unauthorized debits. But if the owner themselves gave the bot access to the account, such a debit is no longer considered unauthorized.
Once you hand over your credentials to the bot, the usual fraud protection mechanisms may not work. Skeptics point to recent incidents. In May, a malicious NFT "hid" instructions that forced the AI to transfer money — this is called "prompt injection." As a result of the attack, about $150,000 was withdrawn from the Bankr wallet associated with Grok. Later, approximately 80% of the stolen funds were recovered. A few weeks later, 14 more wallets on the same platform were affected, and Bankr promised to fully reimburse the losses. So far, there has not been a single proven case where the Grok bot made a mistake with a real bank account.
However, if a crypto wallet is not the target, the damage could be far more serious. The experiment with the Grok bot is partly advertising too: at xAI, they want to integrate the product into everyday financial management. Notably, the very first real mistake will show what Elon Musk's words are worth. The company will be able to respond in two ways: either silently refund all the money, or publicly compensate that very $100 cap.
My take: this is a classic marketing move designed to create hype. The technology of AI agents in finance is still extremely raw, and the legal framework is clearly lagging behind. Relying on verbal promises, even from such a media leader as Musk, is risky. Investors should carefully read user agreements before trusting AI with their hard-earned money.