Google: Demis Hassabis is The head of the AI research unit

Zoom Image

Big thinker: Demis Hassabis to lead AI research at Google in the future

Photograph:

Samuel de Roman / Getty Images

Presumably, few people in the world are better able to assess the opportunities and risks of artificial intelligence (AI) than Demosthenes Hassabis (46). The CEO of Google's DeepMind is one of the absolute pioneers, he strives for the ultimate AI – and yet warns of uncontrollable dangers. At the end of May, he signed the call for global regulation of their technology with other greats such as OpenAI boss Sam Altman (38) – to curb the "risk of extinction" of humanity by AI or nuclear war.

Hassabis, founder of Deep Mind, is one of the leading minds of the scene, the mastermind behind some of the most spectacular successes in technology. In 2016, his AI AlphaGo defeated Lee Sedol, the world's best player in the complex board game Go – the moment is considered a milestone. With his AI Alphafold, which is able to predict the 3D structure of protein molecules, he received the "Breakthrough Prize" in the natural sciences category. His invention caused enthusiasm among physicians: The deeper understanding of proteins makes it easier to understand diseases such as Alzheimer's and accelerates the development of drugs.

Hassabis also sits in one of the most important positions when it comes to the future of technology. As head of Google Deepmind, Hassabis will be responsible for the group's newly merged AI research department. Alphabet CEO Sundar Pichai (50) and Hassabis announced in April that they would merge Deep Mind with Google's AI research division Brain to form the joint unit "Google Deepmind", thus "bringing together two world-class teams", as Pichai writes. The new department has more than 1500 employees. The merger of the two departments to form Google Deepmind is also a reaction of the Alphabet Group to the success of Microsoft's ChatGPT: Alphabet had proclaimed a "Code Red" in the spring to respond to the progress of its competitors in the field of AI and to strengthen its own resources.

With his AI products, Hassabis, of course, wants to improve the lives of billions of people, he writes in a blog post on his company's homepage. And yet, at the same time, he warns of extinction.

Demosthenes Hassabis, whom everyone just calls Demis, is not a man of big words in himself. True, he is named after one of the most famous orators in history – his ancient namesake was Athens' leading statesman in the time of Alexander the Great. But Hassabis usually keeps a low profile, is considered media-shy, and he also rejects interview requests from manager magazin. So who is the man who stands for the emergence and future of AI like hardly anyone else?

From chess prodigy to games programmer to neuroscientist

Hassabis was once considered a child prodigy. At the age of four he learned chess, at 12 he was the second best player in the world in his age group. At first, he wanted to become a professional chess player, traveling around the world and playing tournaments. Until, at the age of 13, he decided that "chess is too small a thing to invest your whole life in".

From then on, the boy devoted himself to his other passion: video game programming. He once said of his first computer: "It seemed like magic to me, like a magical expansion of the brain."

At the age of 17, Hassabis began working at the computer game company Bullfrog Productions after an internship. Peter Molyneux, the co-founder of the company, later recalled the collaboration in an article with the US news magazine "Time": "He is an incredibly competitive person."

After graduating with a bachelor's degree in computer science from Cambridge University in the UK, Hassabis returned to the video game world, developing the video game Theme Park together with Molyneux, among others. In 1998, he became an entrepreneur himself, founded his own video game company, Elixir Studios, but gave it up after a few years to pursue a PhD in neuroscience at University College London.

Entrepreneurship with Deep Mind

After completing his doctorate, in which he focused on the neuronal processes in long-term memory, Hassabis founded his AI start-up: DeepMind Technologies together with Shane Legg and Mustafa Suleyman, in 2010. Hassabis applied the principle of "learning neural networks" to the machine. Inspired by how the brain learns, he trained the software with the help of a computer game and let the AI play millions of times. Through a kind of reward system, the AI learned the optimal strategy to win the game.

In 2013, Google became aware of the company and allegedly prevailed with a purchase price of more than 500 million dollars against Mark Zuckerberg (39), who also wanted to take over the company with his Facebook group. According to media reports, Hassabis chose Google because the company was willing to accept DeepMind's "ethical red lines".

Since 2014, DeepMind has been a subsidiary of Google, but Hassabis continued to fight for autonomy. According to the Wall Street Journal, he spent years trying to establish an independent legal structure. In 2021, however, Alphabet boss Pichai ended these negotiations, according to the Wall Street Journal, citing internal sources.

So now comes the broad solidarity – with Hassabis as the leader of the new AI squad. Hassabis writes in a blog post that he believes this will help him reach his goal of improving people's lives faster with AI products.

AI as a saviour

It is obvious that Hassabis firmly believes in the potential of the technology. In addition to his work at DeepMind, he also runs Isomorphic Lab. He founded the company himself, and it is also a subsidiary of Google. According to the company's website, the company's mission is to use AI to accelerate drug research and find cures for diseases.

In just a few years, Hassabis predicts, a general AI could be developed that is in no way inferior to human intelligence. It would be something like the Holy Grail of technology. In recent years, there has been incredible progress in development. He sees no reason why the pace should slow down, on the contrary, it is likely to become even faster – also due to the further growth in chip and computing power.

Nevertheless, Hassabis is also aware of the dangers of AI, and not just since the spectacular call at the end of May. At the beginning of the year, for example, he told "Time" that one should not act too quickly. The classic motto of the tech industry – act quickly, break the rules and only take care of any mistakes later – is not the right way to go with AI. He now has it partly in his own hands.