🆕 OpenAI's New Image Rules

In today’s email:

🖼️ OpenAI is letting loose with AI images—now you can generate pics of public figures, physical traits, and even some controversial symbols (with rules, of course).
🧾 AI-generated receipts are looking way too real, making refund scams a breeze and trust harder to find.
💭 Need an AI model? From Google’s Gemini to ChatGPT-4o, the options are endless—just don’t expect it to do your taxes… yet.

Intrigued? Keep on scrolling!

🖼️ OpenAI Loosens AI Image Rules

OpenAI just shook up its AI image rules, and things are getting interesting. They’ve decided that AI should be a bit more adventurous—letting users generate images of public figures (hi, Elon and Trump 👋), cultural symbols, and even physical traits that were once off-limits. But with great power comes… well, a whole lot of ethical debates.

What’s New?

• Public figures are now fair game (unless they opt out—like an AI unsubscribe button for your face).

• AI can depict physical traits, but only in neutral/respectful ways.

• Hate symbols? Allowed—but only in educational or cultural contexts (no extremist nonsense).

• Mimic animation styles like Pixar & Studio Ghibli, but not living artists (AI respects copyrights... kinda).

This move is about shifting from blanket bans to smarter, case-by-case moderation. OpenAI swears they’re prioritizing safety, but let’s be real—this could get messy.

🧾 AI Now Creates Fake Receipts

AI can now make fake receipts. Yep, you heard that right. Thanks to OpenAI’s latest update, ChatGPT’s image generation is so good it can whip up restaurant and retail receipts that look totally legit. Scammers are rubbing their hands together while customer service workers everywhere are bracing for impact.

Here’s the breakdown:

• AI-generated receipts look so real you could almost smell the fake burgers.

• They’re not perfect (yet)—some text is a bit wonky, and lighting on the crinkles doesn’t always match. But good luck to the overworked employee spotting those.

• Refund scams incoming—a few tweaks in Photoshop, and boom, free lunch!

• Trust is officially dead—if you thought AI deepfakes were bad, now even a simple piece of paper can’t be trusted.

TechCrunch found even better fake receipts floating around, proving that this cat-and-mouse game is already lost. OpenAI says it’ll crack down on misuse, but let’s be real—the genie’s out, the box is open, and the receipts? Totally phony. Maybe.

💭 Which AI Model Should You Use?

Keeping up with AI models is like trying to drink from a firehose—there are just so many of them! TechCrunch put together a handy guide to the hottest AI models of 2024 and 2025, explaining what they do and how you can actually use them (without getting lost in tech jargon).

A Few Standouts:

• Google Gemini 2.5 Pro Experimental – Great for coding but ironically loses to Claude Sonnet 3.7 in a key coding benchmark.

• ChatGPT-4o Image Generator – The internet went wild when it turned pictures into Studio Ghibli-style anime. Copyright lawyers are sweating.

• Stability AI’s Stable Virtual Camera – Can generate 3D scenes from a single image but still struggles with tricky things like humans and, well, water.

• OpenAI’s Operator – Meant to be your AI-powered personal assistant but, according to one reviewer, it ordered $31 eggs without permission.

• xAI’s Grok 3 – Musk’s latest AI claims to be a math and coding genius. Also, he promised to make it politically neutral (whatever that means).

While some of these models cost a small fortune (looking at you, $200-per-month subscriptions), others are free. AI is evolving fast, and with these models leading the charge, who knows? Maybe one day, your chatbot will do your taxes correctly (but don’t count on it yet).

Other cool AI stuff that is trending right now 🔥🔥

💰 Elon Musk just bought his own company—again—this time merging xAI with X in a $33 billion all-stock deal that promises to blend AI smarts with social media reach but mostly just raises more questions about his ever-growing empire. - Read more

💼 AI-generated job applications are on the rise, and employers worry they might be hiring candidates who can game the system without having the right skills. - Read more

🌍 Google’s AI-powered search feature, Overview AI, has been delayed in most EU countries due to strict regulations. - Read more

🔝 AI chatbot startup StoryFile has made a swift comeback after bankruptcy, launching interactive exhibits at the Museum of the Medal of Honor and the Walmart Museum. - Read more

📞 Krisp is rolling out an AI-powered accent conversion feature, initially allowing Indian English speakers to sound more American on calls. - Read more