Elon Musk is once again shocking the public, this time with a promise of 100% compensation for losses. According to him, if the AI agent Grok, while managing a client's bank account, makes a financial mistake, xAI will fully reimburse the losses. It sounds like a revolution in trust toward artificial intelligence, but upon closer inspection, it is more of a marketing move than a legally binding guarantee.
One Tesla and xAI investor, known by the handle Teslaconomics, decided to test the boldness of this claim in practice. In his post, he asked whether anyone had connected the Grok bot to a real bank account. After all, such an agent, in theory, could track expenses, pay bills, and even flag suspicious charges. However, his interlocutor expressed reasonable concerns: giving AI access to money could turn into a disaster.
Beta test and global plans
The beta version of the product was released on August 11. xAI explained that each agent operates autonomously on its own cloud server, working around the clock, much like a regular user, even while the owner sleeps. This experiment is part of Musk's broader financial strategy. In June, the X Money service with direct transfers between users was already launched, clearly indicating an ambition to integrate AI into everyday capital management.
Legal abyss: the $100 limit
Musk's promise looks like insurance, but the company's documents say otherwise. In xAI's user agreement, it is clearly stated that the results and actions of agents are provided "as is." Moreover, the maximum claim amount is limited to the paid fees or $100—whichever is greater. This is especially ironic given the cost of access: the SuperGrok plan costs $30 per month, totaling $360 per year. The potential damage from an AI error could be orders of magnitude higher.
Musk's response on X does not change these terms. Until xAI formalizes the guarantee in writing, any compensation remains solely at the discretion of the billionaire himself. US banking regulations also complicate the situation. Regulation E protects customers from unauthorized charges, but if the owner themselves gave the bot access to the account, such transactions automatically cease to be considered unauthorized.
Real threats and precedents
Skeptics rightly point to recent incidents. In May, a malicious NFT "hid" instructions that forced an AI to transfer funds—a classic example of "prompt injection." As a result of an attack on the Bankr wallet, linked to Grok, about $150,000 was withdrawn. Later, roughly 80% of the stolen funds were recovered, and after that, 14 more user wallets were affected, with Bankr promising full reimbursement. However, no proven case of a Grok error with a real bank account has been recorded so far.
The experiment with the Grok bot is both advertising and a stress test. The first real mistake will show what Elon Musk's words are worth. The company will either silently return all the money or publicly compensate that very $100, which would be a serious blow to trust in AI finance. From my professional perspective, the market is watching this test with great interest, because its outcome will determine how quickly we trust algorithms not only with cryptocurrency trading but also with managing traditional assets.