2017 was the year that AlphaGo Zero taught itself the game of Go and within 40 days became better than any human or artificial player ever existed. It did so without any human data as input and purely played against itself. As a result, it taught itself strategies and moves no human has ever thought of and arguably progressed the evolution of the game of Go exponentially in a very short timeframe. This achievement marks a significant milestone in the development of artificial intelligence.
In 2018, this will only continue and we will see more examples of artificial intelligence that will behave in unexpected ways, as it already did so this year. In 2017, for example, AI developers from Google built algorithms that had to compete for scarce resources, resulting in increasingly advanced strategies to beat the component. Google Brain developed algorithms that created new encryption methods, unlike any seen before, to protect information from other neural networks. Finally, Facebook had to shut down two algorithms that created its own secret language, unsolicited and used advanced strategies to get what it wanted. If one thing becomes clear from these developments, it is that artificial intelligence will be fundamentally different to human intelligence.
With the AI arms race in full swing, governments and organisations are increasing their investments in the development of ever more intelligent AI. In September 2017, Putin said that “whoever becomes the leader in this sphere will become the ruler of the world”, signalling that Russia will intensify its AI activities. On the other side of the world, China aims to outsmart the USA in AI, with Europe unfortunately nowhere to be seen. The AI arms race seriously scares well-known entrepreneurs such as Elon Musk and Stephen Hawking and a solution for the existential threat of AI is still far away.