AI already dominated chess. Now it is forcing us to play in a different way to continue competing

There is something almost universal in how we understand chess. We imagine it as a duel of pure intelligence, two players in front of a board, trying to anticipate, read the opponent and find the best response at all times. That image still holds true for most of us, whether playing at home or on an app, but in the elite the game has changed a lot. Not because chess has broken down, but because the emergence of increasingly powerful engines has altered the way it is studied, prepared and competed at the highest level.

That change did not come suddenly, although it did leave a very clear scene in 2018. The world championship between Magnus Carlsen and Fabiano Caruana ended with the 12 classic games in drawssomething never seen in the history of the tournament, which dates back to 1886. That result was not an isolated rarity, but the visible sign of a trend that had been maturing for years. The more the best players relied on analysis engines to prepare their games, the narrower the margin to surprise from the start and the more difficult it became to break the balance.

The trick was to stop playing like a machine

To understand what has happened you have to look at how professional chess is studied today. The great masters have been working with engines, programs for years specifically designed to analyze positions and find the best continuations with precision well above the human level. We are not talking so much about conversational models as ChatGPT either Claude, who according to Bloomberg are pretty bad at chessbut from tools like Stockfish or the legacy it left AlphaZero. The underlying issue is that these tools have homogenized part of the knowledge in the elite: many players come to the board with a very similar preparation in the first plays, and that reduces the room for surprise.

This shared preparation began to have a very concrete effect in practice. If both rivals know in advance the strongest lines and the most reliable answers, winning requires much more than avoiding gross errors. For years, the feeling grew that classical chess was becoming more closed at the top, not because of a lack of talent, but because each important detour was much more studied than before. Bloomberg also recalls that the fear of a “death by draw” was not new, but the arrival of engines superior to humans, since Deep Blue in 1997 and later with increasingly stronger domestic programs, gave that fear a different dimension.

Carlsen’s career helps to understand to what extent this change has weighed on the elite. After the 2021 World Cup, an exhausting duel that included an eight-hour game and seven draws, The Norwegian decided not to defend the title again and cited a lack of motivation. He did not abandon classical chessin fact won Norway Chess in 2025 and is still the highest-rated player in the world, but he was showing more and more interest in faster formats such as rapid and blitz, and also in freestyle chess, which alters the initial position of the pieces to neutralize the preparation. The message that this evolution left was quite clear: even the best player on the planet seemed to look for spaces where previous study did not determine everything.

The interesting thing is that the most powerful response came not only by changing the format, but also by changing the way of playing within the board itself. A new generation of grandmasters, already raised with motors, began to assume something that sounds counterintuitive: always following the computer’s first suggestion does not guarantee an advantage over another human. The aforementioned media gives a very concrete example in the 2024 Candidates Tournament, when Praggnanandhaa chose a play considered suboptimal by the engines against Ruy López, took his rival out of preparation and ended up winning.

That’s the key to change. In elite chess it is no longer enough to ask what the best move is in the abstract, it is also important what it is. the most uncomfortable for the person in front of you. Engines may consider several nearly equivalent options, but not all of them create the same type of practical problem on the board. On the other hand, the engine can show you an optimal line, but that does not mean that it has taught you to understand it.

Seen this way, what we are observing is a much more interesting transformation. The engines remain unbeatable and have been far ahead of humans for years, but precisely for this reason they have forced the grandmasters to move the battle to another terrain. Precision continues to be essential, but it is no longer enough on its own if it is not accompanied by judgment, understanding and the ability to adapt.

Images | Florian Cordier | Pavel Danilyuk

In Xataka | A study says that AIs are “cheating” at chess. That’s what we want to think