Nobel Prize winner: AI will revolutionize the world like electrification

At SXSW London, Demis Hassabis answered questions from Bloomberg journalist Francine Lacqua. The man behind Google's DeepMind is certain that artificial intelligence will not only find its way into industry in the future. No area of life will remain unaffected. However, he does not see artificial intelligence as a threat, but as an achievement.

For example, he finds the progress in videos produced entirely by an AI absolutely amazing, but believes that the videos do not appeal to viewers emotionally. Only humans can do that, which is why he does not see AI as a threat to creatives. It is a tried and tested tool for everyday work, but artists must continue to “breathe life” into the videos.

For similar reasons, Hassabis also sees great opportunities for start-ups in AI. Processes would be significantly shortened and simplified by AI, and the point is to use the tools effectively to solve problems.

AI for medicine

In the 2010s, the DeepMind co-founder used AlphaFold to precisely describe the three-dimensional protein structure, solving a 50-year-old problem and winning the Nobel Prize in Chemistry in 2024. With the help of AI, such developments in medicine will be much easier in the future because as soon as a model is available, it can be used by many people. AlphaFold has been open source since 2021 and has also been released for commercial use. In this way, drug development can be shortened from several years to just a few weeks. However, the drugs developed would still have to undergo clinical tests and approval, which is why it will still take some time before the product is ready.

When asked which development he personally hopes for most from AI, the CEO replies that he hopes for a tool for processing emails. Apparently, this activity is a real pain in his everyday life. He would like a system that automatically recognizes the content of general emails and replies to them in his personal style. Google already offers some approaches here, such as summarizing long texts. The Nobel Prize winner did not reveal the extent to which DeepMind is active in this area.

AI is the new electrification

Interestingly, Hassabis believes that AI is currently overrated. In around 10 years' time, however, it will have changed far more than is currently imaginable. In his opinion, work processes will be revolutionized as much as they were in the 1880s with the advent of electrification. In robotics, for example, he still sees a need to catch up, but expects AI to achieve real breakthroughs here in three to five years at the most.

To drive research and development forward, the chemist believes it is essential to understand the technical and physical relationships behind the AI tools. When asked about the recruitment requirements at Google DeepMind, Hassabis was evasive, but it was clear that basic knowledge in this area is very welcome. (uk)