Elon Musk is once again shocking the public, this time with a promise of full compensation for losses if his AI agent Grok makes a financial mistake while managing a client's funds. It sounds like a revolution in trust in artificial intelligence, but upon closer inspection, the legal basis of this initiative looks far from clear-cut.
Experiment with a bank account: a test of strength
One investor, known by the handle Teslaconomics, decided to conduct a demonstrative test. In a post on platform X, he asked whether anyone had connected Grok to a real bank account. The stated functionality is impressive: the agent can track expenses, pay bills, and flag suspicious transactions. While some users see this as the future of personal banking, others, including the investor's own colleagues, express justified concerns that handing over access to finances to an algorithm could lead to disaster.
Musk, unfazed by the criticism, countered: if Grok makes a mistake, xAI will refund the money. However, the beta version of the product, launched on August 11, shows that the founder's statements diverge from the actual terms of use.
Limitation of liability: $100 versus an entire fortune
In the xAI user agreement, it is stated in black and white that the results and actions of agents are provided "as is." The maximum claim amount is limited to the fees actually paid or $100, whichever is greater. Given that access to the bot costs $30 per month (or $360 per year under the SuperGrok plan), it becomes obvious: the stated compensation is a drop in the ocean compared to potential losses in an account.
Moreover, Musk's response on X does not change the legal force of the document. Until the guarantee is put in writing, any compensation remains merely a gentleman's promise, the fulfillment of which depends solely on the billionaire's goodwill.
Risks and regulatory pitfalls
The situation is further complicated by U.S. banking regulations. Regulation E, which protects customers from unauthorized debits, may not apply if the account holder themselves provided access to their data to an AI. In such a case, the transaction would not be considered fraudulent, and the protection mechanisms would prove useless.
There have already been precedents. In May, a malicious NFT "hid" instructions that forced an AI to transfer funds—a so-called "prompt injection." As a result of an attack on the Bankr wallet linked to Grok, approximately $150,000 was drained. Later, about 80% of the stolen funds were recovered. A few weeks later, 14 more wallets were affected, and Bankr promised full reimbursement. But no proven case of Grok making an error with a real bank account has been recorded so far.
The experiment with the bot is, in essence, a marketing move aimed at integrating the product into everyday financial management. The very first real mistake will reveal the true value of Musk's words. The company can either quietly refund all the money or publicly compensate that very $100 stipulated in the contract.
My analysis: Relying on verbal promises in a world where legal documents are decisive is an unforgivable luxury. Until xAI changes the terms of the agreement and provides real insurance, trusting Grok with managing significant assets is nothing short of playing Russian roulette. Investors should remember: marketing statements are not a guarantee, but merely a tool for attracting attention.