Banks are shifting liability for AI agent hacks to customers
Financial institutions are now telling customers that if an AI agent they use gets hacked, the customer bears the risk—not the bank.
According to TechCrunch, more banks are now telling customers that if they use an AI agent to manage accounts and that agent gets breached or misused, the customer—not the bank—is responsible for losses.
This marks a shift in how financial institutions handle AI-related security. Traditionally, banks have been liable when their own systems fail. Now, as AI agents operate on behalf of customers, banks are drawing a line: they provide the connection, but they won't cover damage if the AI tool itself is compromised.
Why it matters for small businesses
If your small business uses AI agents to handle banking, finance, customer service, or other back-office tasks, you need to understand who pays if something goes wrong. Banks are signalling that you own the risk of your AI tools. That means a compromised AI assistant could cost you money with no recourse against your bank.
This also applies beyond banking. As you adopt AI agents for any critical function—managing customer data, handling payments, or processing contracts—you should assume that you are the owner of that risk, not your software provider or financial institution.
What to do
If you use or plan to use AI agents:
- Audit which AI agents you currently use and what data they can access.
- Review the terms of service for each tool to see what liability they accept and what they don't.
- Use strong, unique credentials for any AI agent that connects to your bank or customer systems. Don't reuse passwords.
- Limit what each AI agent can do. If an agent only needs read access, don't give it permission to transfer money or delete records.
- Keep logs of what your AI agents do. If something is wrong, you'll need proof of what happened.
- Consider using your own API keys with AI platforms (WorkBOS supports connecting your own OpenAI, Anthropic, Google, Groq, Mistral, DeepSeek, xAI, Together or OpenRouter keys) so you control the credentials and audit trail.
- If an AI agent handles sensitive data, review your cyber insurance to see whether AI-related breaches are covered.
The trend is clear: AI agents make your business more efficient, but they also put you in the driver's seat for security.
Run it all from one place
Projects, clients, finance, people and AI employees in one system. Start free, no card needed.
Start free →