KM AI: 2020 significant AI milestones

Today, April 20th, in our KM AI course, I gave an introduction to AI terms, using a story-telling approach. Based on Bernard Marr’s piece from 2018, we exposed and better understood the main terms of AI, got to know some thought leaders who shaped the discipline and learned how it all developed to the fabulous achievements we all witness today.

We start back in the 17th century, speaking about the roots, and ended with 2020 achievements. I want to share my thoughts as to the 2020 five top nominations: What was the greatest development of AI in this non-regular year?

The first natural candidate was COVID19. The past year leveraged worldwide connectivity and therefore, ending 2020, we have much more data than we ever had dreamed of. And data, as we all know, is where the wisdom, that AI reveals, lies in. Furthermore, A handful of successful COVID19 AI projects were performed this year, adding to AI reputation and glory.

Another option that crossed my mind was the set of natural language generators. Products like GPT-3 and T-NLG gave a great boost to tools that can create not only chatbot answers, rather complicated, sophisticated texts. Indeed promising.

The third optional milestone was emotional recognition technologies. We are not so amazed already when an AI machine recognizes a cat, however, when it comes to humans’ emotions, understanding what I am saying with no words, through carefully watching my body language is really thrilling!!

I also considered explainable AI. AI started in 1956 with a logical, cognitive approach, and turned 180 degrees towards the end of the century, to a statistical based approach, caring about results, less about reasoning. The new approach indeed brought with it a huge leap (deep learning), but degraded trust of people and organizations in the AI machines’ recommendations. Explainable AI aims to re-add this component of reasoning as an additional layer. It may not bring any breakthrough in technology, yet it will yield a great advance in adoption and organization implementation. Hurray!

And last, but not least… AI democratization. The new sets of tools, whether for cloud services, data to use, AI as a service, algorithmic prepared libraries, AI programming environments and many more, all make it possible, for many more organizations, having less skilled staff, to be able to implement and enjoy this great AI movement. Such is indeed, at my point of view, true democratization of the discipline.

Who is the chosen? I made my choice. But I’ll leave you curious, because it really doesn’t matter. All 5 nominations can surely be the 2020 most significant AI milestone.

Lucky us.