Garry Kasparov on Chess, AI and the Human Role

Garry Kasparov still plays chess almost every day, though he describes it as having fun rather than professional play. Joel Hellermark opens the wider conversation by asking how an authoritarian state might use powerful AI.

Kasparov is skeptical of near-term claims about AGI and superintelligence, describing computing progress as jumps followed by plateaus. Nevertheless, he sees a serious risk when powerful tools are used by hostile governments.

He argues that open democracies are vulnerable to attacks by authoritarian states and that sharing technology cannot be assumed to produce harmony. He sees a turbulent geopolitical competition in which technologies developed in free societies can be turned against them.

Kasparov alleges that DeepSeek benefited from stolen technology and broad data access, and criticizes privacy regulation as a competitive handicap. These are his contested political and policy judgments; he compares models to engines and data to fuel.

His proposed advantage for free societies is the freedom to make mistakes and pursue competing ideas rather than follow central plans. He worries that excessive restrictions could blunt that advantage across what he calls a digital iron curtain.

Hellermark jokes that ChatGPT names Kasparov the greatest chess player. Kasparov says the judgment is subjective, though his record is strong, and acknowledges competition from Magnus Carlsen and Bobby Fischer.

Asked what sustained his success, Kasparov recalls a desire instilled by his mother to make a difference and open new horizons. A win never guaranteed the next win; he tried to analyze even victories without complacency.

For him, the aim was to stay ahead and contribute new ideas, not simply collect titles. He retired from professional chess when that particular drive faded and turned his energy toward human rights, technology and other pursuits.

Hellermark asks whether years of chess permanently shaped his thinking. Kasparov says he also had broad interests, including politics and reading, and that absorbing information and developing ideas felt natural across fields.

He identifies concentration and physical fitness as competitive strengths. Long tournaments could mean preparing in the morning, playing for five or six hours, and analyzing afterward—12 to 14 hours of chess in a day.

After many years the work no longer gave him the same creative satisfaction, so he retired in 2005 at nearly 42. He credits his mother, who raised and managed him after his father's early death, and his wife for supporting different phases of his life.

Kasparov says he invested the recognition earned in chess in a new public life. He values being heard and challenged on subjects beyond the game rather than being viewed only as a former champion.

Machines have transformed chess, he says. Modern engines can defeat even elite players under severe handicaps, including on a phone; they sustain a level of accuracy and intensity humans cannot match indefinitely.

Computer analysis has also enriched the game. Young players can acquire knowledge that took earlier champions decades to accumulate, drawing on both human contributions and machine analysis.

Opening preparation has changed: sharp lines can reward a weaker player who has studied a computer's analysis at home. Kasparov sees a shift toward positional openings that postpone the decisive struggle and test understanding and stamina later in the game.

Hellermark asks about the psychological side of a game with complete information. Kasparov says face-to-face interaction matters: after hundreds of hours opposite Anatoly Karpov, small changes in mood became visible.

He recalls the 1987 Seville championship. After a blunder cost him game 23, he had to win the final game to retain his title; when play was adjourned he was a pawn ahead but had no certain winning plan.

When Karpov arrived for the resumption, Kasparov believed he could see that his opponent had lost confidence in holding the position. Soon afterward Karpov made what Kasparov regards as a decisive positional error.

Kasparov eventually won, though he says Karpov could have tried one last trick that he himself had missed. For him the episode illustrates his mother's saying that a player can lose psychologically before the result is recorded.

Even with all the information on the board, people must decide how much risk to take. Kasparov did not hide his own emotions; he believes the exchange of energy between opponents can affect decision-making.

He recalls coaching the young Magnus Carlsen for about 13 or 14 months after late 2008. Talent alone had taken Carlsen far, and Kasparov helped him make his preparation more systematic.

Kasparov describes Karpov as more positional and himself as more aggressive, while Carlsen combined elements of Karpov and Fischer. Exposure to different ways of evaluating positions, he says, helped Carlsen develop his exceptional talent.

The discussion turns to engines that learn by self-play. Kasparov says serious human-versus-machine competition had effectively ended by 2005 or 2006, before AlphaZero, as engines decisively surpassed top players.

He describes a recurring sequence: machines first seem laughable, then challenge ordinary people, briefly compete with the elite, and finally dominate. The brief period of parity attracts the most attention, even though it is only one phase.

He traces the progression through Belle, Deep Thought and IBM's Deep Blue, alongside engines such as Fritz and Deep Junior. By the early 1990s computers were serious opponents, although they had not yet become totally dominant.

AlphaZero's self-play was a further conceptual shift. Kasparov says it showed that a system could discover strong ideas without inheriting all the assumptions embedded in centuries of human chess knowledge.

Some of those ideas have since entered chess culture. Kasparov sees this as an example of people learning from machine discoveries and incorporating them into their own decisions.

Hellermark asks whether knowledge work will follow chess, where engines can often beat human-plus-engine teams. Kasparov disputes the idea that expert human assistance has become irrelevant.

He argues that AlphaZero has characteristic weaknesses, offering its evaluation of bishops versus knights as an example. A human may not exploit such a weakness over the board but might help another engine target it.

Kasparov believes people can sometimes spot a pattern and adjust after only a few examples, while a learning system may need many iterations. He sees the human role as understanding when and how to fine-tune the system rather than interfering with every step.

Jobs have repeatedly changed with new technology. His advice is to adapt strategies and find applications that combine computing with distinctive human abilities.

Hellermark asks what human quality remains useful to AI systems: intuition, conversation or something else. Kasparov says machines operate inside human culture and institutions, and emphasizes flexibility and the unresolved role of embodiment.

He invokes Joseph Weizenbaum's distinction between computational decisions and human choice, which he associates with wanting something. His claim is philosophical: even if the space for human contribution narrows, a small improvement can be valuable.

Matching the right person, machine and task is crucial, he says. When everyone has access to similar information, even a small improvement in decision quality can provide a significant competitive advantage.

Asked about the meaning of life, Kasparov jokes that ChatGPT is better suited to answering. For himself, he wants to remain relevant and follow a code of honor; he believes the answer differs for each person.

Looking back is painful because he can identify many mistakes. Yet changing any one decision could alter everything that followed, so he prefers to judge the overall balance of a life rather than select isolated counterfactuals.

He recognizes that traits that made him strong at chess may have worked against him elsewhere. On balance, he is proud of what he has accomplished both on and off the board.

Kasparov sees technological progress as mixed: there will be gains and losses, prosperity and destruction. He points to drones used in war, while hoping related technology could eventually help humanity explore other worlds.

He concludes with a political argument that freedom is essential to sustained progress, and that societies should resist systems that suppress individual energy and agency.