- AI makes scams more convincing and gives attackers new targets, like the AI tools businesses connect to their accounts.
- The basics still stop most attacks: multi-factor login, updates, backups and a call-back rule for payment changes.
- Give every AI tool the least access it needs and keep a log of what it does.
Inspired by AI Just Made Your Company Easier to Hack from Alex Lieberman on YouTube. What follows is our own take, written for small business owners.
On Alex Lieberman's show, security researcher David Campbell explains how AI changed hacking. His way of securing AI agents comes down to four checks: how the agent behaves, whose identity it uses, what controls limit it, and whether you can see what it did. Here is what that means for a small business.
Why it matters now
Scam emails used to be easy to spot. AI writes clean, personal messages in seconds, including fake invoices and "new bank details" requests that seem to come from a vendor you know. And every AI tool you connect to email or files is one more door.
Five fixes for this week
- Turn on multi-factor login for email, banking, your website and your domain.
- Use a call-back rule. Confirm any change to payment details by phone, using a number you already have.
- Update and back up. Turn on automatic updates and keep a backup you have actually restored once.
- Limit AI access. Each agent gets its own login with the least access needed, never your admin account.
- Watch the logs. Check weekly what your AI tools sent, changed or deleted.
The FTC's free Cybersecurity for Small Business guide covers these basics in plain language, and the OWASP list of AI risks explains threats like prompt injection.
Right Way Branding builds websites and AI automations with secure logins, limited access and activity logs from day one.



