OpenAI’s high-profile Chief Technology Officer Mira Murati, who has been one of the central figures in the development of ChatGPT and the image generator model DALL-E, said today she’s leaving the artificial intelligence company to do her own “exploration.”

“My six-and-a-half years with the OpenAI team have been an extraordinary privilege,” Murati (pictured) said in a post on X announcing her decision. “There’s never an ideal time to step away from a place one cherishes, yet this moment feels right.”

Murati’s decision to move onto new pastures comes as the AI company reportedly takes significant steps to revamp itself and make it easier to generate revenue and secure funding.

In a report from Reuters today, the company revealed it’s planning to restructure its core business as a for-profit organization, in a move that could deliver more returns to investors. OpenAI, which was first created as a nonprofit research lab before splitting off a for-profit arm, is also said to be engaged in talks with investors over new funding that could see its value grow to $150 billion.

In the wake of Murati’s departure, two more executives also revealed they’re planning to leave the company — OpenAI Chief Research Officer Bob McGrew and research Vice President Barret Zoph.

They become the latest in a string of high-profile executives to leave OpenAI in recent months. In May, the company’s co-founder and Chief Scientist Ilya Sutskever announced he was leaving the firm to start a new AI startup that’s focused on building safer AI for educational purposes. AI safety leader Jan Leike also left the company in May, while another co-founder, John Schulman said last month he’s leaving to join rival AI company Anthropic PBC. Meanwhile, co-founder and President Greg Brockman remains on extended leave.

Murati joined OpenAI back in 2018 from the augmented reality startup Ultraleap Inc., then called Leap Motion Inc., having previously worked at Tesla Inc. With the launch of the original ChatGPT model in late 2022, Murati became one of the most prominent figures at the company as it became a household name, sparking an AI arms race that has accelerated to this day.

In November, Murati’s status was elevated further when she briefly became OpenAI’s interim chief executive officer during the ousting of incumbent Sam Altman. Within a few days, she resumed her role as CTO as Altman negotiated his return to the company. During that week-and-a-half-long debacle, Murati signed an open letter alongside more than 500 other OpenAI employees, threatening to quit the company if Altman wasn’t brought back onboard.

Murati has remained in the spotlight ever since, causing controversy in June when she suggested that AI systems will likely replace some creative jobs. “Some creative jobs maybe will go away, but maybe they shouldn’t have been there in the first place if the content that comes out of it is not very high quality,” she said during The Wall Street Journal’s WSJ Tech Live Conference. “I really believe that using it as a tool for education [and] creativity will expand our intelligence and creativity and imagination.”

As CTO, Murati oversaw a number of advances at the company, including the release in May of its most powerful large language model thus far, GPT-4o, which introduced more human-like, spoken conversational capabilities to ChatGPT. This month, she unveiled the preview of OpenAI o1, also called Strawberry, which is said to have reasoning capabilities that allow it to solve more complex math and science problems.

“Our recent releases of speech-to-speech and OpenAI o1 mark the beginning of a new era of interaction and intelligence,” Murati said in the note announcing her departure. “We didn’t merely build smarter models, we fundamentally changed how AI systems learn and reason through complex problems.”

According to Murati, the reason she’s leaving is to “create time and space to do my own exploration,” though she said she will help ensure a smooth transition to her successor.

OpenAI hasn’t yet said who will take over as its new CTO. In a post on X, Altman hailed Murati’s contributions to the company. “I feel tremendous gratitude towards her for what she has helped us build and accomplish, but I most of all feel personal gratitude towards her for the support and love during all the hard times,” he said. “I am excited for what she’ll do next.” He added that the company will reveal more about its transition plan soon.

Departures prompt leadership reshuffle

As for McGrew and Zoph, they announced their decisions to leave the company a few hours after Murati’s departure was revealed.

Altman said in a second post on X that their departures had nothing to do with Murati leaving, saying they all opted to leave “independently of each other and amicably.” He added that the timing of Murati’s decision meant that it made sense to announce them all at once, in order to facilitate a “smooth handover” to the next generation of leaders at OpenAI.

The company has moved quickly to plug the holes left by McGrew and Zoph, announcing a number of promotions and reshuffles in various aspects of its leadership team. Altman said the company’s vice president of research, Mark Chen, is being promoted to senior vice president of research, where he will lead its research efforts alongside Chief Scientist Jakub Pachocki.

In addition, Matt Knight, who was head of security, is being promoted to the role of chief information security officer. Meanwhile, research scientist Josh Achiam has been given a new position as the company’s head of mission alignment, where he’ll be tasked with overseeing the company’s culture, Altman said.

OpenAI restructures to entice investors

Shortly after Murati announced her departure, Reuters published a report that reveals how OpenAI is working on a plan to restructure its main business into a for-profit corporation. The new business would no longer be controlled by OpenAI’s nonprofit board of directors, the report said, citing people familiar with the matter.

OpenAI’s nonprofit organization would continue to exist, and it would also own a minority stake in the for-profit entity, the sources added. The move is designed to make OpenAI more attractive to investors, and would also have consequences in terms of how it manages AI risks, with a new governance structure, Reuters said.

As part of the plan, Altman will reportedly receive equity in the for-profit portion of the company. That equity could be worth as much as $150 billion, the anonymous sources said.

In response to the report, OpenAI said through a spokesperson that it remains “focused on building AI that benefits everyone, and we’re working with our board to ensure that we’re best positioned to succeed in our mission.”

The report comes a couple of weeks after Bloomberg reported that OpenAI is holding talks with investors over a possible $6.5 billion funding round that would increase its valuation to more than $150 billion. Microsoft Corp., Nvidia Corp., Apple Inc. and Thrive Capital are all said to be involved in the discussions.

OpenAI may need a new cash injection, because it’s faced with some staggering operating costs. According to a recent report by The Information, it has already spent more than $7 billion on model training and an additional $1.5 billion on staffing. ChatGPT alone is said to cost around $700,000 per day in computing resources, while GPT-4’s training costs are said to have exceeded $100 million.

Photo: IBLNews