Mira Murati doesn’t buy the idea that AI should take over jobs. At 36, she’s building systems where humans stay firmly in control. Her startup, Thinking Machines Lab, is based in San Francisco and focuses on AI that enhances human work, not eliminates it. She’s talking to WIRED from her office near Golden Gate Park, where the conversation is less about replacing workers and more about making them better at their jobs.

Murati’s approach isn’t just personal preference. After nearly five years as OpenAI’s CTO, she left in May 2024 during the boardroom chaos that ousted Sam Altman. But she’s not bitter. Instead, she’s doubling down on building AI that serves humans, not the other way around. Her lab’s work centers on systems where AI acts like a supercharged assistant—handling rote tasks so humans can focus on creativity, judgment, and oversight.

At OpenAI, Murati saw firsthand how powerful AI could be. But she also saw how easily it could be misused. “I don’t think we should automate people out of jobs,” she says. “We should automate the boring parts and let people do the thinking.” Her lab’s first product, a tool called CoPilot X, is already in use at GitHub. It doesn’t write code on its own. Instead, it suggests improvements, spots errors, and explains complex parts—all while leaving final decisions to the human programmer.

AI that listens before it acts

Murati’s team isn’t building AI that acts alone. They’re training models to pause, ask for feedback, and adjust based on human input. One project, still in early tests, lets doctors use AI to analyze X-rays. The AI flags potential issues, but it won’t diagnose without a radiologist’s sign-off. Murati calls this “guardrails with guardrails”—systems designed to prevent AI from running wild while still being useful.

Critics argue this approach limits AI’s potential. If systems always need human approval, they’ll move slower. Murati disagrees. “Speed matters, but so does trust,” she says. “If people don’t trust the system, they won’t use it.” She points to early AI tools in hospitals that overwhelmed staff with false alarms. Her lab’s version cuts those distractions by 40% in trial runs. That’s not just a technical win—it’s a trust builder.

The pushback she expects

Murati knows not everyone will cheer her approach. Tech giants like Microsoft and Google are racing to automate as much as possible, often with minimal human oversight. Even some AI ethicists argue that slowing down for “human in the loop” systems wastes time. But Murati’s not backing down. “If we let AI run free without guardrails, we’ll end up with systems that make mistakes we can’t fix,” she says. “That’s not progress.”

Her lab’s next big push is a tool for customer service teams. Instead of replacing call center workers, it’ll handle routine questions but escalate tricky cases to humans. Early tests show it reduces burnout by 25%—because tired employees aren’t stuck answering the same easy calls all day. Murati’s not naive. She knows automation will eliminate some jobs. But she’s betting that, done right, AI can create better ones.

Murati’s work matters because it asks a simple but ignored question: What’s AI for? If the answer is just to cut costs, we’ll get the same old problems in shiny new packages. If the answer is to make life better, we might actually get somewhere. Her lab’s next demo is in three months. Expect something that proves humans still run the show.

What You Need to Know

  • Source: Wired
  • Published: May 15, 2026 at 09:00 UTC
  • Category: Technology
  • Topics: #wired · #tech · #science · #openai · #mira-murati-wants

Read the Full Story

This is a curated summary. For the complete article, original data, quotes and full analysis:

Read the full story on Wired →

All reporting rights belong to the respective author(s) at Wired. GlobalBR News summarizes publicly available content to help readers discover the most relevant global news.

Curated by GlobalBR News · May 15, 2026

🇧🇷 Resumo em Português

O Brasil acorda para uma revolução silenciosa no mundo da inteligência artificial: em vez de substituir empregos, a tecnologia está sendo desenhada para potencializar a criatividade humana. No centro dessa mudança está Mira Murati, engenheira albanesa e ex-CTO da OpenAI, que agora lidera o Thinking Machines Lab com uma missão clara: desenvolver sistemas de IA que atuem como parceiros, não como substitutos.

Murati defende que a IA deve ser uma ferramenta de colaboração, capaz de aumentar a produtividade sem eliminar postos de trabalho. No Brasil, onde setores como o varejo, a saúde e a educação já enfrentam desafios com a automação, sua abordagem chega como um alento — especialmente para profissionais que temem a obsolescência. A relevância de seu trabalho ganha ainda mais peso num país onde a regulamentação da IA ainda engatinha, e onde a discussão sobre ética e inclusão tecnológica é urgente. Com modelos que priorizam o controle humano e a transparência, Murati oferece um modelo que poderia inspirar políticas públicas e empresas brasileiras.

Se sua estratégia vingar, o Brasil pode não só evitar os erros dos modelos anteriores de automação, mas também se tornar um laboratório global para uma IA mais justa e humana.

🇪🇸 Resumen en Español

La polémica sobre si la inteligencia artificial amenaza con arrebatar empleos a los humanos encuentra en Mira Murati, fundadora del laboratorio Thinking Machines y exdirectora técnica de OpenAI, una voz disidente que apuesta por modelos de IA diseñados para potenciar, no sustituir, el trabajo humano. Con su propuesta, Murati desafía el relato más extendido al frente de una corriente que defiende la colaboración entre máquinas y personas como eje de la innovación tecnológica actual.

Murati argumenta que la IA debe actuar como una herramienta de apoyo, no como un competidor laboral, y sus avances en el laboratorio buscan demostrar que es posible crear sistemas capaces de optimizar procesos sin erosionar el papel de los profesionales. Para los hispanohablantes, este enfoque adquiere especial relevancia en un contexto donde la automatización ya genera incertidumbre en sectores como la educación, la sanidad o la industria, y donde la región aspira a posicionarse como líder en tecnología ética. Su trabajo subraya que el futuro no tiene por qué ser una lucha entre humanos y algoritmos, sino una simbiosis que preserve el valor del talento local y la creatividad en español.