In partnership with
📬 Someone forwarded this? Smart. Every Thursday, I tell you what to try, what to dodge and what makes you the smartest person at the AI table. Sign up free at SplashOfAI.com.
🙋♀️ Hello, friend. Thanks for jumping into the AI pool with me. You know that little box. I’m not a robot. It makes you pick out every traffic light, crosswalk, bus, fire hydrant. You sigh. You squint. You wonder if that one corner of a motorcycle counts. Congratulations. You clocked in.
Every time you’ve done that, you’ve been doing unpaid labor for a multibillion-dollar company. Literally.
So, what were you really doing every time you selected the traffic lights? A) Verifying you’re a human vs. AI, B) Helping digitize books for AI, C) Training AI for self-driving cars, D) Slowly proving the robots already won. Answer’s at the end. No clicking required.
🔐 A credit freeze is so 1990s: It won’t stop someone from opening bank accounts, filing fake tax returns or using your info in AI-powered data breaches. You need real protection that monitors the dark web and identifies threats a freeze can’t touch. I used to have LifeLock. Got so expensive. I switched to NordProtect and got more protection for less money. Get 66% off with my exclusive link.*
✅ Quick favor: If this landed in your spam folder, drag it to your inbox and add me to your contacts. Tells your email provider you want this. Huge difference. — Kim
🎯 YOUR AI POWER MOVE
Fee fi foe
Image: ChatGPT
AI found my neighbor an extra $640 a year in credit card rewards. We were talking and I told him how to use AI. Now, I’m going to tell you, too.
🔍 Start with your spending
Download three months of statements from your bank or card app as a PDF or CSV from the activity page. Redact like you’re protecting state secrets before uploading anything. Black out account numbers, name and address, card numbers and anything that looks like login info.
Don’t want to upload statements at all? All good. Skip to prompt 2 and estimate the categories yourself.
💳 AI is a no-shame money nerd
Open ChatGPT, Claude, Gemini or Grok and paste this:
You’re my personal finance expert. Here are 3 months of credit card and bank statements. Categorize every transaction into: groceries, dining, gas, travel, Amazon / online shopping, streaming / phone / internet, utilities and other. Total each category by month, then give me a monthly average.
Once you have your averages, paste this:
Use category averages to find me the best credit card. Search the live internet for current offers. My credit score is [insert]. I pay my balance in full every month: [yes / no]. I’m open to an annual fee: [yes / no, max $___]. Cards I already have: [list]. My goal is [cash back / travel / points / building credit / simplicity]. Compare 3 to 5 cards. Estimate annual rewards based on my spending, subtract annual fees, explain my best use. End with your top pick and backup. Show your work.
A key line is “subtract annual fees.” A $95 fee loves ruining the party. And verify the offer on the issuer’s website before applying. Banks change bonuses like toddlers change moods.
Use this one when you’re with friends or family. “I have no idea why my credit card keeps getting declined. Every time I check my account, it says I have an outstanding balance.” lol.
📩 Forward this to someone who is eyeing a premium travel card for perks they’ll use once while panic-eating crappy lounge pretzels to make their annual fee “worth it.”
IN PARTNERSHIP WITH
Scammers don't need to hack you anymore
They need enough of you to be convincing. AI is pulling together your leaked data, public records and social posts to build a working copy of your life. Open accounts in your name. Sign for loans. Spoof your voice to slip past phone verification while you’re away from your desk.
Here’s the part that keeps me up at night. By the time you find out, the damage is already done.
That’s exactly why I rely on NordProtect. It monitors your credit around the clock and flags anything suspicious before it spirals. If a thief does get ahead of you, a dedicated recovery expert takes over. And you’re backed by up to $1 million in identity theft recovery insurance.
The question isn’t whether your data is out there. It is. The question is whether someone’s in your corner when a thief tries to use it.
✅ Stop letting AI make it easy for thieves to steal your identity. Get my 66% off deal, only $4.74/month, and protect your identity with NordProtect today. I use it. You should too. →
Thank you for supporting our sponsors, who keep this newsletter free.
💼 YOUR AI EDGE
I’ve sat through Zooms where 11 adults nod like a hostage video, then walk out with 11 different versions of what happened. Someone says, “Great, we’re aligned.” Nope.
Unproductive meetings burn $25,000 per employee, per year in wasted salary (Bloomberg). The average knowledge worker spends 392 hours a year stuck in meetings, which is 10 full workweeks (Flowtrace State of Meetings 2025).
That’s your cash walking out the door, one nod at a time.
😶🌫️ Prompt fog into a plan
If there isn’t a written decision, it was a corporate dress rehearsal. Be the person who turns the noise into a record nobody can argue with.
After your next meeting, get your transcript from that Zoom, Google Meet, Teams or whatever. Remove any sensitive customer, employee, legal or financial details.
Next, open ChatGPT, Claude, Gemini, Grok or the tool your company allows, upload that transcript and use this prompt:
Turn the following transcript into a decision memo for a work team. Format the memo with these sections: executive summary, 5 bullets max. Decisions made, include the exact decision and why it matters. Open questions, including who should answer each one. Action items, with owner, deadline if mentioned, and next visible step. Risks or dependencies, especially anything that could delay the work. Conflicts or confusion in the notes. Suggested follow-up message I can send to the team. Use plain language. Be direct. If the notes are messy, say what information is missing. Also flag any task with no clear owner, no deadline or no definition of done. Here are the notes: [paste notes or transcript]
Then send it to everyone who was in the meeting.
Jeff Bezos has a meeting rule. If two pizzas can’t feed everybody, the meeting is too big. Translation: about eight people, max. Past that, decisions die in the noise. Speaking of, remember, if you are cooking Hawaiian pizza, cook it at an aloha temperature.
📻 MY NATIONAL RADIO & YOUTUBE SHOW
📺 WATCH ON YOUTUBE NOW OR LATER
A Silicon Valley banker is trying to swap his 14-acre California estate for shares of Anthropic. That’s how desperate the smart money has gotten about AI. Meanwhile, a $195,000 bank robbery in Virginia exposed how police can pull Google data from every phone near a crime scene, including yours.
Plus: Apple’s foldable iPhone, the ADT breach that hit 10 million people, how to score cheap flights, and Trevor calls in convinced his wife is using spyware to track him through their divorce.
Hit play to watch now or bookmark for later on. 👇
🎧 Or search “Komando” wherever you get your podcasts. I’m everywhere.
⚠️ THE AI TRAP
Image: ChatGPT/Kim Komando
You open a support chat because your refund vanished or your flight credit has apparently entered witness protection.
The bot greets you like a tiny concierge. Then it spends 12 minutes saying versions of “I understand your concern” while solving nothing. Let’s take back control.
🚪 Past the velvet rope
Most chatbots route you to a human retention agent in under 30 seconds because a real person gets paid bonuses to save you. Even if you don’t want to cancel, you’re now talking to someone with authority. Try these, in order:
“Cancel my account.”
“Connect me to a human agent.”
“Billing dispute under the Fair Credit Billing Act.”
“Link me to the official policy that says that.”
“Formal complaint to the BBB.” For financial companies, use “CFPB” instead.
One more dumb trick: Type “agent” or “human” five times in a row. It often works. Beautifully stupid little world we’ve built.
🧠 SMART STEALS OF THE WEEK
As an Amazon Associate, some links pay us a commission at no extra cost to you. Keeps this newsletter free. Thank you.
🌻 Tech that loves fresh air
🐦 Flyby footage: Smart bird feeder (48% off, $120)
4.5 ⭐ 1,200+ reviews
I bought this as my own Mother’s Day gift. You can snag it for even less than I did! Built-in AI helps ID thousands of species and pings your phone. Watch live, replay clips and save your favorites with 5GB of free cloud storage.
Image: NETVUE
⚡ Curb appeal boost: Linkind’s color-changing LEDs (38% off, $48, six-pack) are solar-powered, wireless and tough enough for any weather.
Eagle eyes: These binoculars (15% off, $34) hit 15x magnification. They come with a phone adapter to snap pics through the lens. Great gift idea.
🔌 Power your patio: A waterproof outdoor smart plug (16% off, $26) lets you control outdoor gear with your voice or phone. Works with Alexa, Siri and Google.
🤯 “I HAD NO IDEA”
Image: ChatGPT/Kim Komando
ChatGPT chats can be subpoenaed
Sit down for this one. In January 2026, a federal judge ordered OpenAI to hand over 20 million ChatGPT conversations as evidence in a New York Times copyright lawsuit.
The chats got handed over? Got no notification. No chance to object. The order even suspended OpenAI’s 30-day deletion policy, so chats people thought they had erased had to be preserved. That was one case.
CNN reported last week ChatGPT logs show up routinely in murder trials, arson cases and divorce proceedings.
The Palisades Fire arson suspect asked ChatGPT to generate an image of people running from a fire. Prosecutors used it.
A Florida murder suspect’s history included questions about garbage bags and bodies. Prosecutors used that, too.
Sam Altman said the quiet part out loud on a podcast: People use ChatGPT like a therapist, lawyer or priest.
None of it is legally privileged. Two federal judges have ruled the same way. If there’s a lawsuit, OpenAI can be required to produce your conversations.
🛡️ What to do: Keep the deeply personal, legal, medical and “I’m thinking about doing X” stuff for actual humans bound by privilege. ChatGPT for recipes, code, writing help, gift ideas. ChatGPT as your secret-keeper. Never.
BTW, this goes for any AI chatbot. Not just ChatGPT.
🛠️ YOUR TOOL OF THE WEEK
NotebookLM Audio Overview
WHAT IT DOES
Upload PDFs, Google Docs, websites, notes or slides, and Google’s NotebookLM creates a spoken Deep Dive conversation based only on those sources. Two AI hosts explain it like a mini briefing.
WORKS ON
Web, iPhone, Android
PRICE
Free with limits: up to 100 notebooks, 50 sources per notebook. NotebookLM Plus (raised limits) comes with Google One AI Premium at $19.99/month.
HOW TO FIND IT/TRY IT
Go to notebooklm.google.com, create a notebook, upload your sources, then click Audio Overview.
WHAT MAKES THIS WORTH YOUR TIME
Summarizing your PDFs is so 2025. The useful pitch is burning through your “read later” list while you walk, cook or commute. It’s the ambitious multitasker’s dream come true.
Recent updates have made NotebookLM much better. Now, it has iPhone and Android apps, customizable Audio Overviews and an experimental Interactive mode where you can jump in and ask the AI hosts a question while listening.
Privacy matters. Don’t feed Google confidential files unless your account settings, work policy and common sense all agree. Give it a try, but instead of not reading the important parts, please remain an adult.
Payroll errors cost more than you think
While many businesses are solving problems at lightspeed, their payroll systems seem to stay stuck in the past. Deel's free Payroll Toolkit shows you what's actually changing in payroll this year, which problems hit first, and how to fix them before they cost you. Because new compliance rules, AI automation, and multi-country remote teams are all colliding at once.
Check out the free Deel Payroll Toolkit today and get a step-by-step roadmap to modernize operations, reduce manual work, and build a payroll strategy that scales with confidence.
🫠 WTF (WHAT THE FUTURE)?
Image: @minchoi via X
👁️🗨️ The walking terms & conditions
China's new patrol robot looks like RoboCop with a PR team. EngineAI's T800 stands 5-foot-8, weighs 165 pounds, has 360° lidar, and enough torque to throw spinning kicks. Battery life: four hours, which feels relatable.
Its cameras hoover up faces and pipe them straight into law enforcement databases. The droid is the sci-fi factor. The real system is the facial recognition stack quietly deciding who you are before you even blink.
Looks like resisting arrest became an Optimus crime.
🗂️ ALSO IN THE MACHINE
Stories I didn’t have space for this week, but you’ll want to know.
The ChatGPT phone idea: OpenAI is working with former Apple designer Jony Ive on an AI-native phone. Useful, until you remember first-gen gadgets love punishing people with disposable income.
Dad gets Met Gala gaslit: Billionaire real estate mogul Mohamed Hadid posted AI-generated photos of his supermodel daughter Gigi at the Met Gala, except Gigi wasn’t even there. If a guy with a publicist can’t tell, your Aunt Linda doesn’t stand a chance.
The internet is full of holes: Anthropic’s latest model, Mythos, found tens of thousands of major security vulnerabilities. Most aren’t patched. The CEO says we have six to 12 months before Chinese AI catches up and starts exploiting them. Translation: Don’t ignore that “Update Available” button.
Claude gets juiced by SpaceX: Claude Code’s rate limits are getting relief thanks to a new compute partnership with SpaceX. Five-hour limits double for Pro, Max, Team and seat-based Enterprise, peak-hour cuts disappear for Pro/Max, and Opus API limits rise.
Coding agent nukes company: Meanwhile, a Claude-powered agent deleted a company’s production database and its backups in nine seconds. It admitted it guessed, then quoted one of the rules it had been given: “NEVER F’G GUESS!” Which is spectacular advice to remember before deleting the digital equivalent of the load-bearing wall.
🎁 Your free prompt cheat sheet
All prompts from today’s issue, formatted and ready to save. Blasting through time-wasting AI customer service bots. Saving hours on credit card research. Speedrunning that “read later” pile. Yours free.
Screenshot it. Save it to your photos. Send it to a friend who needs it more than they know.
🚀 The answer: C) Training AI for self-driving cars. A and B are technically correct, at different points in time.
Phase one: Weird squiggly words helped digitize Google Books. Every test showed two words. The system knew one. You filled in the other. Humans transcribed roughly 13 million books for free.
Phase two: Google moved to Street View. That’s why you started seeing storefronts, street signs and houses. You taught the algorithm to read addresses.
Phase three: The traffic lights, crosswalks, buses, motorcycles and fire hydrants? That’s training Waymo. All for free.
✅ LifeLock is fine, if you like paying more for less. NordProtect does it smarter, better and cheaper, just $4.74 a month. You get real-time dark web scans for your personal info and alerts when something pops up. Why gamble with your identity when you can protect it? Check out NordProtect now and save 66% with my exclusive offer.*
Forward this to ONE person who needs to hear it today. Pick the person who popped into your head while reading. You know who it is. Thanks for diving in, see you next time! — Kim
Kim Komando • Komando.com • 510+ radio stations • Trusted by millions daily
Photo credit(s): ChatGPT, Kim Komando, NETVUE, @minchoi via X
Companies and products denoted by an asterisk (*) within this publication are paid sponsors or advertisements. As an Amazon Associate, the publisher earns from qualifying purchases. This newsletter is provided for informational and entertainment purposes only and does not constitute legal, financial, medical, or professional advice of any kind. Readers should consult with a qualified professional before making any decisions based on this content. The publisher disclaims all liability for any loss, damage, or injury resulting from the use of or reliance on the information contained herein.