The Woman Who Helped Build ChatGPT Has a New Vision for AI

The Woman Who Helped Build ChatGPT Has a New Vision for AI

If you've ever worried that artificial intelligence is on a path to replace you — your job, your judgment, your sense of purpose — there's a conversation happening right now in Silicon Valley that you should know about. And it involves one of the most credible voices in the entire AI industry.

Mira Murati spent years as the Chief Technology Officer of OpenAI, the company behind ChatGPT. She was, by most accounts, one of the key people responsible for turning that technology into something millions of people actually use. Then she left to start her own company, called Thinking Machines Lab. And the direction she's heading is surprisingly different from where much of the industry seems to be going.^[1]

In a recent interview, Murati said something that doesn't get said enough in AI circles: she isn't interested in building systems that automate people out of their jobs. What she wants to build, instead, is AI that genuinely collaborates with humans — where the person stays in control and the machine supports their thinking rather than replacing it.

Why this matters more than it might seem.

Right now, a lot of the biggest AI projects are racing toward what insiders call "full autonomy" — systems that can independently research, decide, and act without any human involvement. Some companies are openly celebrating the idea of AI that will do the work of entire teams of people. The message, sometimes spoken out loud, is that human workers are an obstacle to efficiency.

Murati is pushing back against that framing, and she has the credibility to make people listen.^[1] When someone who helped create one of the world's most powerful AI systems says "keep humans in the loop" — meaning, keep people involved in the decision-making process rather than sidelining them — it carries real weight.

Think about what this looks like in real life. Consider a small business owner named Sandra who runs a bookkeeping service for local contractors. She could use AI tools to handle routine invoice sorting, flag unusual expenses, or draft client reports. But Sandra's clients trust her judgment — her ability to notice when something feels off, to ask the right question, to make a phone call when the numbers tell a story that software might miss. An AI that assists Sandra is valuable. An AI that tries to be Sandra is a different thing entirely, and probably a worse outcome for her clients.

Or think about a retired nurse who now volunteers doing medical translation for immigrant families at a community clinic. AI could help her quickly look up unfamiliar terms, prepare for appointments, or summarize a patient's records. But the trust in that room, the cultural sensitivity, the human connection — that's hers. The goal should be to make her more effective, not to replace her.

The broader picture is messier, of course.

At the same moment Murati is articulating this more human-centered vision, the rest of the industry is anything but unified. Elon Musk and Sam Altman — two of the most prominent figures in AI — are literally suing each other in federal court, with a jury now deciding a case that has, according to observers, managed to make both men look bad in the process.^[2] Meanwhile, major tech employers like Cisco are cutting thousands of jobs specifically to redirect that money toward AI investment.^[3] The gap between the optimistic vision and the day-to-day reality for workers is real and growing.

That tension is exactly why Murati's position is worth paying attention to. She's not a philosopher or a critic writing from the outside. She's a builder — someone who understands technically what these systems can and can't do — and she's making a deliberate choice to build toward collaboration rather than replacement.

What should you take from all of this?

The AI industry is not a single thing moving in a single direction. There are genuine disagreements among the people building these tools about what they should do and who they should serve. Murati's voice represents one side of that argument — the side that believes the most valuable AI is the kind that makes you better at being you, not the kind that makes you unnecessary.

That's not a naive position. It's actually a pretty good business strategy too, since people tend to trust and keep using tools that feel helpful rather than threatening. But more importantly, it's the kind of AI development that most of us — workers, small business owners, retirees, caregivers — would actually want to live with.

The question is whether that vision wins out.

Sources