DeepMind co-founder Suleiman leaves Google: management style has been questioned

On January 22, it was reported that Mustafa Suleyman, co-founder of DeepMind, a subsidiary of Google's artificial intelligence (AI), has left Google to join venture capital firm Greylock Partners.

Over the past 7 years, London-based DeepMind has made breakthroughs in artificial intelligence, but it has struggled to ensure that its technology is used "ethically" by its Google.

DeepMind was founded in the UK in 2010, and Suleiman was one of the three founders of DeepMind. In 2014, DeepMind was acquired by Google for $500 million. Now, Suleiman has left the firm to become a partner at Greylock Partners, a Silicon Valley venture capital firm. Previously, Suleiman has sparked controversy, including employee complaints about his aggressive management style.

After being acquired, DeepMind tried to keep a distance from Google, secure a research climate that insulated it from commercial pressure, and tried to devise a governance structure for its powerful AI technology. DeepMind's first attempt was to create an independent AI ethics committee to oversee its work, but with little success. Later, DeepMind tried to create a legal framework to give it more independence within Google, but without success.

In a transcript of a conversation with Greylock partner Reid Hoffman, published yesterday, Suleiman said: "We have made so many mistakes in the way we try to form ethics committees, and I'm not sure I can say it was absolutely successful. But I do think radical experimentation is essential here. Moreover, we have also tried different oversight committees, different ethical charters, and different types of research. ”

Speaking about the tech industry's broader efforts in this area, Suleiman said there are still many unanswered issues around technology platforms, software, and, of course, artificial intelligence. In fact, Google itself is struggling with "how to introduce checks and balances in the development of artificial intelligence."

Google's plans to set up an independent AI ethics committee were dashed by employee protests against the committee's composition. In addition, two of Google's ai ethics executives also left the company.

In 2019, after DeepMind employees complained about its management style, Suleiman was forced to leave DeepMind and move into Google's headquarters as vice president of AI products and policy (losing some of his previous responsibilities). Speaking about employee complaints, he said: "I really messed up. I was demanding and quite ruthless. I set some rather unreasonable expectations, which led to some people working in a very harsh environment. I am very sorry about that. ”

In 2016, DeepMind's artificial intelligence program AlphaGo defeated world champion and South Korean Go master Lee Sedol, and DeepMind has been famous ever since. In 2020, DeepMind used artificial intelligence to predict how proteins would fold into different shapes, hailing a breakthrough that could be useful in areas such as drug discovery.