Did You Know: James Manyika Published His First AI Paper as a Student in Harare, Long Before He Ran Research at Google

When Google wants to explain what artificial intelligence is for, the person it usually sends is James Manyika. He oversees Google Research and Google Labs, sits on the company's senior leadership team, and co-chaired the United Nations Secretary-General's advisory body on AI.^[1][2]

His very first piece of published work was an AI paper too. He wrote it in Harare.

A neural networks paper from 1992

"The first thing I ever published in my whole life was actually a paper on AI in 1992, as an undergraduate," Manyika said in 2024. "It was actually on modeling and training neural networks. This is in Zimbabwe, which also surprises people sometimes, given the timeframe."^[3]

The timeframe is the remarkable part. Neural networks are the idea behind today's chatbots and image generators, but in the early 1990s most computer scientists had written them off. Manyika was studying engineering at the University of Zimbabwe when he met a graduate student from Toronto who was working on artificial intelligence, and through that student learned about the research of Geoffrey Hinton, who would later be called the godfather of AI. Manyika was hooked.^[4]

When he took the idea to Oxford, his advisers warned him not to mention neural networks, because no one would take him seriously.^[4]

From Mbare to Prince Edward to Oxford

Manyika was born in 1965 and grew up in what was then Salisbury.^[1] He has described a childhood in Mbare, in a segregated city, during the war years. What gave him his vision as a young man, he said, was his father, who went to the United States on a Fulbright fellowship and visited Cape Canaveral. The young Manyika decided he wanted to be an astronaut.^[5]

After independence in 1980 he was one of the first Black pupils to attend Prince Edward School, and he has recalled travelling there under police escort.^[4]

He went on to study electrical engineering at the University of Zimbabwe, won a Rhodes Scholarship, and at Oxford earned a master's in mathematics and computer science and then a doctorate in AI and robotics. He became a research fellow at Balliol College and co-wrote a technical book on how robots combine information from many sensors, published in 1994.^[1][6][7]

He never made it to space, but he got close to the people who do. He spent time as a visiting scientist at NASA's Jet Propulsion Laboratory, where by his own account he worked in a group building machine learning systems for the Mars Pathfinder programme.^[3][8]

The long detour

Then came what he calls a long detour.^[3] Manyika joined McKinsey, the consulting firm, in the United States in 1997 and stayed for about 25 years, 13 of them as chairman of the McKinsey Global Institute, its research arm. He advised technology companies, wrote widely on automation and the future of work, and was appointed by President Barack Obama as vice chair of the White House's Global Development Council.^[8][1][2]

He kept one foot in AI throughout. He has said he was involved early on with DeepMind, the London lab founded by Demis Hassabis that Google later bought.^[3]

In 2022 Google created a new post for him, senior vice president of technology and society, reporting directly to chief executive Sundar Pichai. A year later the job was widened to take in Google Research and Google Labs.^[1] Scientific American interviewed him in June this year about how AI is changing the way science itself gets done.^[9]

Optimist with a warning label

Manyika is not a simple cheerleader. He argues AI will bring enormous benefits, yet he has also signed a public statement, alongside hundreds of researchers, saying that the risk from AI should be treated as seriously as pandemics and nuclear war.^[4] It is also worth keeping in mind that a senior Google executive is not a neutral voice on whether Google's products are good for the world.

Where he is most persuasive is on who gets left out. Around a visit to Accra last year, he warned that without reliable electricity, internet access and training, "the digital divide risks becoming an AI divide".^[10] A point that other influential africans have noted in the tech industry, without the basic infrastructure, innovation is stunted.

Why his story lands differently here

Africa is personal for him, he has said, because he grew up in Zimbabwe, where he began working in computing and first realised AI's immense potential to address local challenges.^[10]

It is easy to read a career like his as a story about leaving. The more useful reading is about what was possible before he left. A student at the University of Zimbabwe in the early 1990s, with none of today's computing power, was able to do original work in a field that the world's leading universities were ignoring. The ingredients were a chance meeting with someone who knew the field, and a stubborn interest.

That is worth remembering now that Econet says it has hired more than 100 engineers for its AI unit. His advice to students in Ghana applies equally in Harare, Bulawayo or Gweru: be ambitious, because Africa has talent, energy and creativity in abundance, so study AI, launch projects and become entrepreneurs.^[10]

He also had a caution for anyone who thinks learning to code is the whole job. Coding is essential, he said, but mathematics, statistics and an understanding of systems and data are the foundation. His worry is that too many people think coding alone is enough, and it isn't.^[10]