In partnership with

Want to get the most out of ChatGPT?

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide

Josh here. This one is a little shocking… Can you believe it?

OpenAI Just Killed Its Best Use Cases (And That's the Point)

Here's a wild one: OpenAI just updated its usage policies to explicitly ban the exact things their AI is demonstrably good at. No more medical image analysis. No tailored legal or medical advice without a licensed professional involved. Just "general information" and a polite redirect to go see a real doctor.

This is fucking bizarre when you consider the research. We've seen studies where AI outperforms radiologists at detecting certain cancers. GPT-4 has helped people navigate complex legal challenges. The technology works for this stuff. And now OpenAI is saying: yeah, but you can't use it that way.

Here's what's really happening.

OpenAI isn't making a technical decision—they're making a liability decision. The policy update reads like it was written by lawyers who just got out of a very long meeting about what happens when someone's AI chatbot misdiagnoses cancer or gives someone advice that lands them in legal trouble.

Look at the specifics: "provision of tailored advice that requires a license, such as legal or medical advice, without appropriate involvement by a licensed professional." That's not about capability—it's about ass-covering. They're drawing a bright line between "general information" (safe) and "tailored advice" (lawsuit waiting to happen).

The thing is, this reveals something deeper about how AI is actually going to integrate into regulated professions. We keep hearing breathless takes about AI replacing doctors and lawyers. But OpenAI—the company with the most to gain from that narrative—just said "nah, we're good."

Why does this matter?

Because it shows us the real friction point isn't technological, it's institutional. The AI can read your CT scan. But who's liable when it's wrong? The AI can draft your legal brief. But who gets disbarred when it cites fake cases?

OpenAI is basically saying: we built a powerful tool, but we're not willing to absorb the risk of people using it to make high-stakes decisions without professional oversight. Which is probably smart! But it also means the "AI revolution in healthcare" you keep reading about is going to look more like "AI assists doctor who is ultimately responsible" rather than "AI replaces doctor."

The facial recognition ban is interesting too—no databases "without data subject consent," no real-time biometric ID in public spaces. That's OpenAI explicitly refusing to enable surveillance capitalism, even though it's clearly a profitable use case.

The pattern here: OpenAI is drawing lines that prioritize institutional risk management over maximizing use cases. They're choosing not to be the company that enables unlicensed medical practice or mass surveillance, even if the technology could do it.

It's almost quaint—a tech company voluntarily limiting how people can use its product. But it also tells us something about how the AI future actually unfolds: slower, more cautious, and way more intertwined with existing professional gatekeepers than the hype would suggest.

The AI can see the tumor. You just can't rely on it to tell you what to do about it. That distinction matters more than the technology itself.