Claude or ChatGPT? A values question, not a features one

Say you want a simple answer. Should you use Claude or ChatGPT? You have a chart in mind, one column against the other.

The more useful move is to turn it into a different question. What matters to you about the company you hand your thinking to every day? Most people have not thought about it that way. So let me walk you through how I think about this, one person to one person.

Most comparisons of these two tools are about speed, price, and benchmarks. That is fine, and it is not this piece. This one is about something quieter. Who builds the tool, and what do they believe. If you are a purpose-led expert, that question is not soft. It is the whole thing.

Does it matter who builds my AI?

Yes, it matters more than most people expect. An AI tool is not a neutral pipe. Its training choices, its guardrails, its willingness to say no, and its pace of release all carry the priorities of the company behind it. When you use one every day, you are living inside those priorities. So it is worth knowing what they are.

I think about it like choosing a bank, or a supplier, or a co-author. The output can be great and the values behind it can still sit wrong with you. You are allowed to care about both. In fact, if your own work is built on trust, you probably should.

What does Anthropic believe?

Anthropic, the company that makes Claude, was founded in 2021 by a group of people who left OpenAI, led by Dario Amodei, partly over disagreements about how carefully powerful AI should be developed. Safety is not a side project for them. It is the reason the company exists, at least by its own account.

Two structural details tell you a lot. Anthropic is a public benefit corporation, which legally binds it to weigh public good alongside profit. And it created a Long-Term Benefit Trust, a group of financially disinterested people who can elect and remove a growing share of the board over time. As of 2026, trust-appointed directors are a majority of that board. Structure is not a promise, but it is harder to walk away from than a slogan.

On the product side, two ideas show up in how Claude behaves. Constitutional AI trains the model against a written set of principles it consults during training. The Responsible Scaling Policy sets safety thresholds that are supposed to gate how far the company pushes its models. Amodei also writes at length in public, including his essay Machines of Loving Grace, where he is openly hopeful about what AI could do for health and science while staying blunt about the risks. You can read the man in his own words. I find that reassuring.

What about OpenAI?

OpenAI, the company behind ChatGPT, has a genuine mission and a huge record. Its charter, from 2018, says its goal is to make sure artificial general intelligence benefits all of humanity, with benefits broadly distributed. ChatGPT put useful AI into the hands of hundreds of millions of people faster than anyone thought possible. That is a real contribution, and I will not pretend otherwise.

The debate around OpenAI is about how it has changed. It started as a nonprofit, then added a capped-profit arm in 2019, and in October 2025 restructured again so that its for-profit side is now a public benefit corporation, the OpenAI Group, overseen by the nonprofit OpenAI Foundation. As part of that change, the old cap on investor returns was removed. Supporters see a company raising the capital it needs to compete. Critics see mission drift. Both are looking at the same facts.

Two events stay in my mind. In November 2023, the board briefly removed Sam Altman as CEO, saying he had not been consistently candid with them, and after an employee revolt he returned within days with a new board. And in 2024, the leaders of OpenAI's superalignment safety team left. One of them, Jan Leike, wrote publicly that safety culture and processes had taken a backseat to shiny products. I take him at his word, and I also know one person's account is not the whole story.

Over the past years, safety culture and processes have taken a backseat to shiny products.

So which should I use?

Use the one whose builders you trust with your work, and know that either choice can be a good one. I am not going to hand you a single right answer, because there is not one. This is my informed opinion, not a fact about the world.

Here is mine. I work with Claude. I do human-first work with people who are the business, whose name is on the door, whose reputation is the product. For that work I want a maker whose structure and stated priorities put care ahead of pace. Anthropic reads that way to me. The public benefit corporation, the trust, the safety frameworks, and Amodei writing openly about the risks all point in a direction I can stand behind when I put a tool in a client's hands.

None of that makes me right for you. If ChatGPT fits your work, if you trust OpenAI's direction, if the features you need live there, use it with a clear conscience. Plenty of people I respect do. The point is not the logo. The point is that you chose on purpose.

How do I choose for myself?

Start with your own values, then read the makers against them. A few questions I use:

• What is this company legally bound to prioritize, and who holds it to that?
• How does it talk about risk when the news is not good? Openly, or only in press releases?
• When people who cared about safety left, what did they say on the way out?
• Does the tool's behavior in my own use match what the company claims to value?

Then test both for a week on your real tasks. Values matter, and so does whether the thing helps you do good work. When the two line up, you have your answer. Try both, and pick the one that feels like yours. That is all I want for you.