Good morning! Today we're taking a look at what's really behind some of the political ads you saw before this week's elections, plus two more things worth knowing about AI.
In today's issue: why some of the political ads you watched before Election Day may have been made by AI, a free trick for finding subscriptions you forgot you were paying for, and a warning about AI reading your own social media posts to write you a personalized scam email.
Let's get started.
NEWS
That Political Ad You Saw Last Week? It May Have Been Made by AI
If you watched any campaign ads before this week's elections on November 3, there's a good chance at least one was partly made by artificial intelligence — and you probably had no way to know.
A nonpartisan research group called the Wesleyan Media Project tracked political ads this election cycle and found campaigns spent more than $137 million on 220 video ads that were at least partly AI-generated. Compare that to the last general election, in 2024, when the same group found campaigns spent under $50,000 on AI ads. Republican campaigns and their support groups were responsible for 81% of these ads and 88% of the spending, though the researchers note campaigns of both parties used the technology.
Here's the part that should give you pause: more than 70% of these AI-made ads included no disclosure at all telling viewers that AI was used. One example: in a Texas Senate race, an ad for one candidate showed an AI-generated version of his opponent writing in a Bible with a crayon while delivering a speech the opponent never actually gave. The real candidate said it was a deepfake and pointed out he's actually pushed for laws banning this kind of deceptive AI video.
Takeaway: Don't assume a political ad shows something that really happened, especially if a moment in it seems unusually dramatic or out of character for the person shown — AI can now fabricate a believable clip, and most ads using it aren't required to say so.
— Fortune
TRY THIS
A Free Way to Find the Subscriptions You Forgot You're Paying For
Most of us have at least one subscription we're still paying for and barely use — a streaming service, an app, even a magazine. Free AI chatbots like ChatGPT or Gemini can help you sort through them in a few minutes.
Start by jotting down every subscription you can think of — streaming, apps, meal kits, memberships, whatever comes to mind. Then type or say to ChatGPT or Gemini: "Help me identify which subscriptions I should cancel or downgrade," and list what you wrote down. The AI will help sort them into ones you use often, ones you use occasionally, and ones you'd honestly forgotten you had — then suggest which to cancel, pause, or switch to a cheaper plan.
It won't automatically see your bank statement or cancel anything for you — you still have to do that part yourself, through each service's app or website. But it's a free, quick way to get organized before you go hunting through old statements.
Takeaway: Grab five minutes, list what you're subscribed to, and ask free ChatGPT or Gemini to help you sort the "use it" pile from the "forgot I had it" pile — it could save you real money every month.
STAYING SAFE
Did You Just Post Vacation Photos? A Scammer's AI Might Be Watching
Researchers at the University of Texas at Arlington found that AI tools can read your public social media posts — your vacation photos, the events you're attending, even things you're worried about — and automatically write a scam email tailored just for you. For example, if you post photos from a hotel stay, you might soon get a very convincing fake email offering a "discount" from that exact hotel chain.
In the researchers' tests, people who read these AI-written emails rated them as highly believable, specifically because they felt relevant and personal rather than like generic spam. Even more concerning: the safety features built into the AI tools that are supposed to block this kind of misuse only caught the attempt between 0% and 19% of the time.
The research, led by professor Shirin Nilizadeh, will be presented at a cybersecurity conference in 2027, but the techniques it describes are already possible today.
Takeaway: Consider waiting until you're back home to post vacation or event photos, keep your old posts private when you can, and for any "surprise" offer or discount email, go directly to the company's real website or call a number you look up yourself — never click the link or call the number in the email.
Until tomorrow,
EZ AI