A while ago, I had the pleasure of becoming immersed in the future of AI in Boston, a field where the boundaries between human control and autonomous technology are increasingly blurring. The most impressive experience was at Boston Dynamics, the maker of robots like Atlas and the dog-like Spot. It was the first time my company managed to visit them, and it left a deep impression on me.

We have all seen Atlas grow up from a clumsy robot - standing unsteadily on its feet ten years ago and barely able to open a door - to an extremely agile and fast humanoid that can effortlessly run up stairs and perform elegant somersaults. Despite how advanced these devices have become, they are still mostly controlled and operated by a human using a remote control. 

But this could change quickly. Boston Dynamics believes in a breakthrough for 'autonomous robots', within this decade. And the evolutions in robotics have certainly been progressing rapidly in recent months. Elon Musk's TeslaBot is perhaps one of the best-known robots alongside those from Boston Dynamics. But there are many more: Apollo from Apptronik, Neo from the Norwegian 1X (in which OpenAI recently invested), Phoenix from Sanctuary, or the robots from Figure.

All these companies believe that we can build robots that can extinguish raging fires and rescue people from burning buildings. Robots that autonomously inspect nuclear power plants to detect leaks or check sewers. 

In other words, machines that take on 'dull, dangerous and dirty work,' the jobs that people no longer want to do.

We primarily associate 'autonomous' with self-driving cars. These have proven to be much harder to realize than we expected. Only two companies were granted permission last summer to deploy self-driving cars on public roads in San Francisco. One of them (Cruise) had to suspend its public activities because one of its cars had dragged along a victim in an accident.

This immediately brings up the fundamental dilemma in AI ethics: the trolley problem. If the brakes of an autonomous car fail, do you let the algorithm decide to run over an old woman or two children? This is an unsolvable ethical dilemma.

The world of autonomous is greatly being accelerated by military developments. Drones have become an extremely handy war tool, from Ukraine to the Middle East. But most drones are still First Person View: a human controls them remotely. The problem is that there are too few drone pilots. So both Russia and Ukraine are experimenting extensively with autonomous navigation and target recognition. They are building killer AI drones. 

The ethical dilemmas suddenly become very different.

The integration of AI into military weapon systems will cause a seismic shift. AI-driven autonomous drones, robots, or 'cyberattacks' can completely change warfare. And they also have the potential to cause a quantum leap in the science of AI and 'autonomous.'

Boston Dynamics' very first prototypes are somewhat sadly displayed in the entrance hall. But almost everywhere on the signs next to them, DARPA is mentioned. This stands for the American Defense Advanced Research Project Agency, established in 1958 after the Russians humiliated the US with their Sputnik satellite. Out of nowhere, research for 'national security' received an unlimited budget. The results were accordingly: the internet, for example, was a DARPA project, as was the GPS. And now the autonomous robots.

"War. What is it good for?" According to Edwin Starr's 1970 hit song, the answer is 'Absolutely nothing.' But unfortunately, that's not true. Millennia of 'civilization' show a chilling correlation between innovation and warfare. The old Cold War gave us space travel, computers, and the internet. Today, it is undeniable that AI has become a geopolitical weapon. It will only be developed more intensely due to the current increase in war zones.

Europe, however, does not have a DARPA. We do not have a hyper-coordinated military agency, and certainly no strong link between scientific innovation and our warfare capacity. The gap between Europe and the tech superpowers around us could significantly widen due to the current acceleration in autonomous warfare.