Democratisation of artificial intelligence

Artificial intelligence has been the long-term dream of humankind for decades. Already 70 years ago, visionary thinkers had an abstract idea of an electronic brain able to perform human-level tasks tirelessly and without pay. However, until today this has been a goal somewhere in the future. During the last five years, progress has been immense, and the breakthrough of human-level AI is here, now.

Artificial intelligence can be roughly categorised into weak AI and strong AI. The former consists of systems that solve narrow, specific tasks but can’t generalise to other domains. The license plate recognition systems of Finnish company Visy represent a good example— the computer reads your car’s number plate while entering a parking lot or other access-controlled facilities. However, it will never learn other tasks, such as replying to your pile of emails or washing your dishes. Generalisation over problem domains is the fundamental characteristics of strong AI: Something that a child can do, but machines not just yet.

Nevertheless, the weak form of AI is expanding without any limits. The mobile phone can now recognise your speech or propose that a journey home from work would take 11 minutes right now. Behind the hype and excitement in the news, there is also a less widely known movement of making AI available to everyone — the democratisation of artificial intelligence. Indeed, big companies are following the lead of academia and openly publishing their tools in open source. This, together with affordable computation platforms such as the Amazon AWS, allows almost everyone to become a data scientist.

Finland is in a good position to contribute to the new technologies. The ecosystem of skilled labour, universities and companies is already here. People are the critical factor in the equation, and we are very pleased to see the number of students doubling every year for our machine learning courses at Tampere University of Technology.