The AI risks that matter in real life: scams, deepfakes, bad information, and privacy mistakes.
AI tools can be useful and still create risk. This page is not here to scare you off. It is here to help you keep your guard up while you use them.
With less than 30 seconds of audio, AI can now clone a person's voice convincingly enough to fool family members. Scammers use this for "grandparent scams" — calling elderly relatives pretending to be a grandchild in trouble and urgently needing money. Deepfake video technology is becoming almost as easy.
Business version: scammers impersonate the owner or a senior employee over phone or video call to authorize wire transfers or change payment information.
Establish a family code word for emergencies. Always verify unexpected requests by calling back on a known number — not the one that called you. No legitimate emergency requires an irreversible wire transfer in the next hour.
Phishing used to be easier to spot because the writing was sloppy and the setup was obvious. AI has improved the scammer's side of the job. The messages now use your real business name, cleaner language, and details pulled from public information.
Common targets: fake invoices, fake package delivery issues, fake bank security alerts, fake vendor payment updates.
Any email that asks you to click a link and enter credentials or change payment information should be verified via a separate channel — call your bank directly, log into your account manually (not through the link). Enable multi-factor authentication on every account that matters.
AI can generate convincing fake websites — including fake versions of your business — in minutes. Scammers create fake contractor sites with AI-generated reviews and photos, collect deposits, then disappear. If someone finds a "CitrusAIworks" that isn't this site, it might be a scam operation using a similar name.
Search your business name regularly to see what comes up. If a fake site appears, report it to Google. Customers: always verify a business has a real address and reviews on Google before paying deposits.
AI language models generate text by predicting what should come next — not by looking up verified facts. When they don't "know" something, they don't say "I don't know." They generate a plausible-sounding answer. This is called a hallucination, and it can be catastrophic if you act on it.
Real examples: AI citing court cases that don't exist, inventing product specifications, providing incorrect legal or medical information, generating fake statistics that sound credible.
Never publish AI-generated content without fact-checking specific claims. Never rely on AI for legal, medical, financial, or regulatory information without verification by a qualified professional. Use AI for drafts and ideas — use verified sources for facts.
Free AI tools — including ChatGPT's free tier — may use your conversations to improve their models. That means sensitive information you type could potentially become part of future training data. Employees who use public AI tools for work without official oversight are called "shadow AI" users, and they can unknowingly expose customer data, trade secrets, or proprietary information.
Never enter customer data, social security numbers, financial records, passwords, or confidential business information into free AI tools. Check the privacy settings — ChatGPT lets you turn off training data usage in Settings. For business use with sensitive data, use paid enterprise plans that offer stronger data protections.
The scam tactics are old even if the tools are new: urgency, impersonation, and pressure to send money or hand over credentials. Slow down and verify through a channel you trust.
For the tools themselves, treat AI output like a draft, not a fact source. Keep sensitive data out of free tools and turn on multi-factor authentication everywhere you can.