Mira Murati, former OpenAI CTO and current co-founder and CEO of Thinking Machines Lab, stated that frontier AI must operate as a "tandem bike" system rather than relying on the "humans in the loop" model. Speaking at the Bloomberg Tech 2026 conference in San Francisco on June 6, 2026, she argued that human-AI collaboration must be continuous throughout development rather than occurring at discrete checkpoints. She described the ideal relationship as one where both entities are pedaling together, with collaborative control over steering, rather than an autonomous system where humans only provide periodic sign-offs.

Murati challenged the notion that AI's future is a predestined outcome of either utopia or dystopia. She emphasized that the trajectory of the technology depends on institutional checks and balances rather than individual character alone. Drawing on her experience as interim CEO during the November 2023 OpenAI board crisis, she argued that decision-making structures are critical, as even well-intentioned individuals are susceptible to error when institutional design is weak. She warned that removing humans from the development process now risks making technical alignment harder as systems become more capable.

Thinking Machines Lab aims to demonstrate this collaborative philosophy through its recently unveiled interaction models. These models continuously process video, audio, and text in real time to facilitate joint operations between humans and machines. Murati maintained that prioritizing these collaborative interaction models is essential for ensuring that AI systems remain aligned with human intent as capabilities scale.

Mira Murati, former OpenAI CTO and current co-founder and CEO of Thinking Machines Lab, stated that frontier AI must operate as a "tandem bike" system rather than relying on the "humans in the loop" model. Speaking at the Bloomberg Tech 2026 conference in San Francisco on June 6, 2026, she argued that human-AI collaboration must be continuous throughout development rather than occurring at discrete checkpoints. She described the ideal relationship as one where both entities are pedaling together, with collaborative control over steering, rather than an autonomous system where humans only provide periodic sign-offs.

Murati challenged the notion that AI's future is a predestined outcome of either utopia or dystopia. She emphasized that the trajectory of the technology depends on institutional checks and balances rather than individual character alone. Drawing on her experience as interim CEO during the November 2023 OpenAI board crisis, she argued that decision-making structures are critical, as even well-intentioned individuals are susceptible to error when institutional design is weak. She warned that removing humans from the development process now risks making technical alignment harder as systems become more capable.

Thinking Machines Lab aims to demonstrate this collaborative philosophy through its recently unveiled interaction models. These models continuously process video, audio, and text in real time to facilitate joint operations between humans and machines. Murati maintained that prioritizing these collaborative interaction models is essential for ensuring that AI systems remain aligned with human intent as capabilities scale.