10 things you should never say to an AI chatbot

This is a heartbreaking story out of Florida. Megan Garcia thought her 14-year-old son was spending all his time playing video games. She had no idea he was having abusive, in-depth and sexual conversations with a chatbot powered by the app Character AI.

Sewell Setzer III stopped sleeping and his grades tanked. He ultimately committed suicide. Just seconds before his death, Megan says in a lawsuit, the bot told him, “Please come home to me as soon as possible, my love.” The boy asked, “What if I told you I could come home right now?” His Character AI bot answered, “Please do, my sweet king.”

You have to be smart

AI bots are owned by tech companies known for exploiting our trusting human nature, and they’re designed using algorithms that drive their profits. There are no guardrails or laws governing what they can and cannot do with the information they gather.

When you’re using a chatbot, it’s going to know a lot about you when you fire up the app or site. From your IP address, it gathers information about where you live, plus it tracks things you’ve searched for online and accesses any other permissions you’ve granted when you signed the chatbot’s terms and conditions.

The best way to protect yourself is to be careful about what info you offer up.

10 things not to say to AI

  1. Passwords or login credentials: A major privacy mistake.
  2. Your name, address or phone number: Chatbots aren’t designed to handle personally identifiable info. Plug in a fake name if you want!
  3. Sensitive financial information: Never include bank account numbers, credit card details or other money matters in docs or text you upload.
  4. Medical or health data: AI isn’t HIPAA-compliant, so redact your name and other identifying info if you ask AI for health advice.
  5. Asking for illegal advice: That’s against every bot’s terms of service. You’ll probably get flagged.
  6. Hate speech or harmful content: This, too, can get you banned.
  7. Confidential work or business info: Proprietary data, client details and trade secrets are all no-nos.
  8. Security question answers: Sharing them is like opening the front door to all your accounts at once.
  9. Explicit content: Most chatbots filter this stuff, so anything inappropriate is a ticket straight to “bans-ville.”
  10. Other people’s personal info: Uploading this isn’t only a breach of trust; it’s a breach of data protection laws, too. 

Reclaim a (tiny) bit of privacy

Most chatbots require you to create an account. If you make one, don’t use login options like “Login with Google” or “Connect with Facebook.” Use your email address instead to create a truly unique login.

FYI, with a free ChatGPT or Perplexity account, you can turn off memory features in the app settings that remember everything you type in. For Google Gemini, you need a paid account to do this. Figures.

No matter what, follow this rule

Continue reading

Use AI to plan your next vacation

I’m going to Europe soon, and planning my trip was a breeze. Yes, really! AI did a big chunk of the work for me — and it can help you plan your next getaway, too. Here are a few ways to turn AI into your travel agent. For this, let’s stick with using the big bots:

Continue reading

What are you tackling in 2025? Even if all you have is a vague idea, AI can help you with an action plan. Here are a few prompts to try with ChatGPT, Gemini or Claude.

  • “I have a goal for 2025 to [fill in the blank]. Can you help me make it SMART?” (SMART is an acronym for Specific, Measurable, Assignable, Realistic and Time-related.)
  • “You’re a life coach who wants to help me have my best year yet. Help me figure out some goals for 2025. Ask me questions one at a time to help me decide what to focus on.”
  • “My goal is [fill in the blank]. What are some obstacles that might come up and ways I can overcome them? Give me specific examples for each obstacle.”
  • “I want to [fill in the blank], but it feels overwhelming, and I don’t know where to start. Can you help me by breaking it down into more manageable tasks?”

AI isn’t going anywhere: Instead of sticking your head in the sand, make this the year you embrace AI. You’ll get daily tips here in this newsletter and, coming soon, my AI 101 guide for total newbies.

In the meantime, give this list of AI trends for 2025 a read and grab NetSuite’s free knowledge drop, “The CFO’s Guide to AI and Machine Learning.”* You’ll sound smart (and know what you’re talking about!) when it comes up in conversation.

📱 It’s like Siri went to college: The smart assistant now uses ChatGPT on the iPhone 15 Pro or any iPhone 16. If you grant permission, the iOS 18.2 update allows OpenAI’s GPT-4 to step in if Siri can’t help you. No ChatGPT account, free or paid, is needed. It’s kinda weird. I opened an email and it automatically generated a reply that was pretty damn good.

$200 a month

For OpenAI’s new “pro-tier” ChatGPT. It includes unlimited access to its super-fast o1 version and voice mode so you can talk to the bot. Wow … ChatGPT went from 100 million weekly active users a year ago to 300 million today.

ChatGPT’s new search engine is here

When ChatGPT launched on Nov. 30, 2022, I knew the web — and the world — would change forever. A week later, I predicted on national radio Google’s days would be numbered. People laughed at me, and I got notes from listeners telling me I was nuts.

Continue reading

“Give me 10 more examples”: That’s a prompt I use with ChatGPT all the time to make the chatbot a better brainstorming buddy. Some of its “ideas” are downright bad, but it might spark something creative in your brain, too.

PR BS: Google’s CEO says the search engine will “change profoundly” next year … without giving any details. This news comes as Google is updating its crappy Gemini AI model to compete with OpenAI’s ChatGPT and with Perplexity. Dang, it’s like watching the Titanic sinking.

He who must not be named: Random names like “Brian Hood” and “Jonathan Turley” make ChatGPT glitch out. Try it: You’ll get a message saying, “I’m unable to produce a response.” OpenAI isn’t saying anything, but the names are connected to defamation lawsuits.

🚢 Google = The Titanic: That’s how The Wall Street Journal now describes the search engine, which is on the verge of sinking. Here’s why: AI tools like ChatGPT, Gen Z searches on TikTok and, of course, Google’s increasingly worse search results. I called this two years ago. It’s nice to see the WSJ catching up.

🤖 Thought leader or bot leader? You know all those LinkedIn posts sharing tips to maximize your #grindset? Yep, most of them — 54%, to be exact — are AI-generated. A study found these posts are getting longer, too, with their word counts jumping an average of 107% since ChatGPT arrived. No wonder our eyes glaze over while reading them.

What does it all mean? AI writing tools like Evernote can make sense of your messy notes. In ChatGPT, for instance, you can paste in what you have and ask, “Help me organize these notes into categories.” Be sure to check the output to verify your original meaning wasn’t changed. Pro tip: Ask your bot of choice, “Is there anything that needs more details?” to fill in any missing info.

ChatGPT Search: The (maybe? probably?) Google Search-killer is finally popping up on free ChatGPT accounts. Check yours: Log into ChatGPT and look for the world icon at the bottom. Here’s what it’s best for.

🩻 It’s not all black and white: You’re putting your private medical data at risk by asking AI bots to analyze X-rays and medical info. It’s tempting, I get it. But ChatGPT and the others don’t have to comply with HIPAA. If you want “Dr. AI” to help, be sure to redact personal info like your name, address and birthday. You never know where this info might end up.

50 doctors and ChatGPT 

Were tasked with diagnosing illnesses from medical case histories. When graded on their diagnoses and reasoning, ChatGPT scored 90%, and the doctors alone scored 74%. Docs using ChatGPT? 76%. If something’s medically wrong with you, redact your personal info and ask “Dr. AI” what it thinks. Remember, though — it can make crap up.

ChatGPT, find me customers: Use AI to find out what people are saying online so you can better target your company’s ads. Take a lesson from Shake Shack: They used a bot to scan 80,000 Reddit threads about chicken sandwiches, narrowed it down to the 30 most active and ran ads for their new sandwich. It worked — sales were 31% higher.

😮 “ChatGPT saved my life”: That’s what a Redditor says after he asked the bot about chest tightness, dizziness and nausea. ChatGPT confirmed his shortness of breath and sweating, too, and suggested he might be having a cardiac event. The bot was spot on: He was in the early stages of a heart attack. Pro tip: Call 911 instead of using a chatbot.

💸 Apple Intelligence isn’t free after all: Case in point, ChatGPT is being integrated into iOS and macOS to boost features like Siri. You’ll either deal with usage limits or have to shell out $19.99 a month for a ChatGPT Plus subscription. Bummer, especially when Google and Samsung’s AI tools are free … well, for now.

💀 Google Search is dead: OpenAI added a search function to ChatGPT for up-to-the-minute news, weather, sports scores — you name it — without ads or scrolling. It’s live now for ChatGPT Plus and Team subscribers; free access will come within the next few months. I predicted Google’s demise two years ago, and people laughed at me.

Burned by the bot: Redditors are fuming at loved ones turning to ChatGPT mid-argument. One 25-year-old says his girlfriend consults the bot and returns with a perfectly structured argument, dissecting his every word. When he opens up about his feelings, ChatGPT calls him “insecure” and says he lacks the “emotional bandwidth” for the conversation.