Elon Musk is encouraging users to let Grok Bot shop on their behalf, including by providing payment information, even as SpaceXAI has updated its privacy policy to place responsibility on users for the information and actions they provide to the AI system.
Musk has promoted Grok Bot as an agent capable of searching for products, identifying deals, adding items to shopping baskets and completing purchases. In one example, he said he asked Grok Bot to research Lego Star Wars kits for young children and order them.
The shopping capability may require users to provide payment details. Cybersecurity researchers have raised concerns about the risks of sharing financial information with an AI system that can act autonomously, particularly if the agent makes mistakes, exposes data or is manipulated by malicious instructions.
FCRF Launches CP-FRM to Build India’s Next Generation of Fraud Risk Professionals
SpaceXAI Privacy Policy Shifts Responsibility to Users
SpaceXAI recently updated its privacy policy, removing earlier language that warned users against sharing certain personal information in prompts and inputs.
The revised policy states that the company cannot control what users provide and that users are responsible for actions involving payment information as though they had taken those actions themselves.
The wording differs from Musk’s earlier assurance that users would be compensated if Grok Bot made a mistake.
SpaceXAI continues to discourage users from submitting sensitive personal information involving areas such as race, ethnicity, political orientation, health and biometric data. Credit card information is not specifically listed in that category in the policy excerpt, although financial information can carry significant risks if exposed or misused.
The policy also states that personal information submitted through files, images, audio, voice, video or other inputs may be used by Grok when generating responses. It notes that personal information included in inputs may also appear in outputs.
Researchers Warn of Data Exposure and Transaction Risks
Cybernews senior researcher Aras Nazarovas warned that a shared AI system could potentially mix up context and unintentionally reveal financial information to other users or use stored payment information in unintended transactions.
Concerns over financial data exposure have already surfaced in relation to Grok Bot. Cybernews reported a case in which a technology executive’s personal financial information was allegedly exposed to employees through Slack.
Information security researcher Rasa Jurgutytė said encouraging users to place greater trust in Grok for financial transactions could make the system more attractive to attackers.
She warned that prompt injection and similar techniques could potentially be used to manipulate an AI agent handling financial actions. Such attacks are designed to influence an AI system through malicious instructions embedded in content it processes.
Researchers also cautioned that AI agents can make errors even without malicious interference. A shopping agent acting autonomously could misunderstand instructions, purchase the wrong item or carry out an unintended transaction.
Autonomous Shopping Could Create New Security Challenges
The growing use of AI agents for shopping introduces questions about how financial data is handled and who bears responsibility when automated transactions go wrong.
If users provide images of credit cards or other payment credentials to an AI assistant, the information may become part of the data processed by the system. That creates potential exposure if the information is reproduced in outputs, incorrectly associated with another interaction or accessed through a security weakness.
The ability of AI agents to browse websites, compare products and complete purchases also increases the importance of protecting them against manipulation. Attackers could attempt to influence an agent into revealing information, choosing a malicious seller or making an unintended purchase.
For users, the convenience of automated shopping therefore comes with additional security considerations. The latest SpaceXAI privacy-policy language makes clear that responsibility for information supplied to Grok and actions involving payment details may rest with the user, even as the company promotes broader use of its AI agent for autonomous shopping.
About the author — Suvedita Nath is a science student with a growing interest in cybercrime and digital safety. She writes on online activity, cyber threats, and technology-driven risks. Her work focuses on clarity, accuracy, and public awareness.
Follow for daily updates on cybercrime, corporate fraud, DFIR, hacking, investigations, and digital forensics
