In the world of AI, we have no Mendeleev table, no overview. It is not a world of slow, deliberate change, but one of hasty leaps forward. When I recently took some time to go through the things that came my way when I moved out of my parents’ house, I came across an old acquaintance: a large poster with Mendeleev’s table from the 1970s. In my youth, this table was not an abstract school object, but a daily presence. My father, an honorary professor at KU Leuven, taught chemistry and made sure that this colorful arrangement of the elements was prominently displayed in our home. For our family, the table was almost a natural piece of decor, a silent witness to the order in the matter that makes up the world.
I looked at that table with nostalgia. Whereas as a child I mainly saw the colors and the strict structure, now I see the patience and tranquility it exudes. The table is more than a diagram; it tells the story of an academic who, in the nineteenth century, attempted to organise the jumble of elements by arranging them according to their properties and atomic mass. He deliberately left empty spaces, convinced that nature had an order, even if humans did not yet fully understand it. Later, other scientists filled in precisely those empty spaces. His table was therefore not only an overview, but also a guide for the future.
Whereas the table brought order to the complexity of atomic matter, in the complex world of AI there is a constant tension between innovation and confusion. Out of curiosity, I opened my children’s course books that day to see how many subjects had been added in the meantime, 50 years later. I looked for new additions, each of which was the result of in-depth research and international patience. It takes years of experimentation and verification before an atom is given its official place in the table. One laboratory is never enough; only after independent confirmation and strict control by the responsible unions does a discovery gain legitimacy.The contrast with my current field of work, artificial intelligence, could hardly be greater. In AI, we have no table, no overview, no fixed vocabulary. Our world is driven by chaos, speed and fragmentation. New terms, concepts and models appear almost daily, often without anyone agreeing on their exact meaning. This is not a world of slowly considered change, but one of hasty leaps forward, in which no one knows exactly which boxes still need to be filled in – and sometimes not even what we are actually looking for. Mendeleev left empty boxes open because he was certain that nature, in collaboration with intelligent humans, would one day fill them.
In AI, we have no table, no overview, no fixed vocabulary. Our world is driven by chaos, speed and fragmentation. New terms, concepts and models appear almost daily, often without anyone agreeing on their exact meaning. This is not a world of slow, deliberate change, but one of hasty leaps forward, in which no one knows exactly which boxes still need to be filled in – and sometimes not even what we are actually looking for. Mendeleev left empty boxes open because he was certain that nature, in collaboration with intelligent humans, would one day fill them. In AI, there are certainly still important empty boxes to fill, but the insights surrounding these boxes are rather vague.
Companies launch products that reach millions of users before there has been any independent testing. Papers appear in open repositories and are picked up within days, often without peer review. And because there is no neutral committee to determine what is sustainable, researchers and companies fill the gap with self-made benchmarks that mainly confirm their own progress. Whereas the table brought order to the complexity of atomic matter, in the complex world of artificial intelligence there is a constant tension between innovation and confusion.
Looking back at that large poster in our house, I understand better why my father so often gazed at it with admiration. Mendeleev’s table is not only a classification of matter, but also a reflection of a culture in which inertia and evidence are the highest values. My father hoped that I would follow his passion for chemistry, but perhaps that is precisely why, as a rebellious teenager, I chose a different path. Today, after years in the storm of AI, I look at my decision back then with different eyes. As I look ahead to 2026, I long for “a Mendeleev’s table for AI” – a global initiative that brings order to this chaos.
Geertrui Mieke De Ketelaere is AI-expert en adjunct professor at the Vlerick Business School. She wrote several books and is a great keynote . Mieke was elected ICT Person of the Year in 2024
