Chess has often been called the “fruit fly” of artificial intelligence because it’s the perfect subject for experimenting on. That has never been truer than at the U.S. Chess Championships, where the presence of the sly-eyed young internet cheat, Hans Niemann, is providing an interesting clinical investigation of the vague term “artificial performance enhancement” and what it really ought to mean.
How do we draw the line between sports cheaters and sports strivers?
The national tournament in St. Louis is a fascinating observational study: Can Niemann play chess live and in person at a high enough level, while under surveillance from silicon scanners and full body wandings of his backside, to convince observers that he’s a purely human cerebral genius? The 19-year-old has been accused by world champion Magnus Carlsen of using synthetic intelligence to play chess and lying about it, a claim partly verified by a devastating report from Chess.com, which has banned Niemann for “likely” cheating in more than 100 online games. Niemann protests that he only committed a couple of youthful indiscretions online — and “never, ever” for money — and has offered to play naked if it helps prove his merit as a grandmaster. Through the fifth round in St. Louis, he has a win and three draws to go with a loss to grandmaster Fabiano Caruana, not quite what you would expect from the next genius but not a definitively guilty performance according to statistical models, either.
Where to draw the line between acceptable striving and “dirty,” immoral or “unnatural” striving? Compared to a grandmaster who consults a software engine to win at the board, Barry Bonds looks like a Victorian. For years, Bonds and other athletes of the steroid era — including my old friend and co-author, Lance Armstrong — were considered the standard for unscrupulous striving by anti-doping bureaucracy for their use of pharmacological substances to build muscle, physically recover or level the playing field. But the emergence of the Niemann dilemma and the juxtaposition of cheating in chess with doping in baseball as Aaron Judge’s home run pursuit revived old concerns about Bonds’s single-season record have clarified a basic problem. In all our anxiety over artificialities, we haven’t thought carefully enough about how to distinguish between rank cheating and the murkier world of performance enhancement.
A great baseball slugger who juices is not trying to minimize his effort — he’s trying to maximize. Sports dopers are many things, but they aren’t lazy. They’re the opposite. They’re excessively driven. Whereas a chess player who relies on AI to solve a board problem is not seeking to maximize but to minimize. He’s putting out the least possible amount of effort. It’s an utterly flaccid thing to do. That kind of cheating leads to atrophy, not enhancement.
The fear that athletes will become too synthetic, little more than a set of muscle mechanics that exhibits some humanlike traits, is a gross oversimplification. Athletes are more than just well-built machines. Their constitutions are systems, certainly, governed by chemistry and physics. Move an arm in a certain way enough times and it will improve its functionality. But that’s not a complete explanation for them. Something is left out. There’s a blank in our understanding. How athletes translate mere physicality into phenomenal performance and tremendous plasticity — how they are so astoundingly improvable — is a beautiful mystery of temporal processing. The anti-doping movement has been entirely grounded in an obsession with the physical. But what makes someone great — whether in baseball or chess — is a complex intersection of work, intentionality, incentive, opportunity, sensory perception, insight, psychology, economics and umpteen other factors.
What really elevates performance? We don’t entirely grasp how a great athlete emerges “from the biological wet-ware of 100 billion neurons connected by 100 trillion synapses,” to borrow a description from Stanford professor Surya Ganguli in his essay “The intertwined quest for understanding biological intelligence and creating artificial intelligence.” An athlete presents scientists with a profound “credit assignment problem,” Ganguli writes. Suppose a tennis player misplays a ball. “Which one of your 100 trillion synapses are to blame?” Ganguli asks. This is why no AI can (yet) truly mimic the stupendous neuronal-synaptic orchestra that is Bonds at the plate or Stephen Curry in motion toward the basket. To do so, AI would have to “put together in some sense parts of the computer scientist, neurobiologist, psychologist, and mathematical theorist in the same brain,” Ganguli writes.
The anti-doping movement dumbs down this mystery to single cause, and it equally fails to get at the motives behind sports performance enhancement. A trove of studies shows that athletes who are inclined to juice are driven not by an impulse to shortcut but by “perfectionism.” The clinical world defines perfectionism as a multidimensional tendency toward “striving for flawlessness.” It’s therefore nonsense to talk to such athletes about the “morality” of performance enhancement when to them the greater immorality is to leave any potential wasted. Good luck discouraging perfectionists from experimenting with their bodies. You might as well tell an astronaut not to go too far from the farm.
Ultimately, cheating in chess and so-called “doping” in sports raise the same question: What is the purpose of the contest? Most broadly, games are for learning, aren’t they? The great champion Garry Kasparov once said that “chess is life in miniature,” by which he meant the ultimate lesson of the chess board is how to make difficult decisions and accept the consequences with a certain amount of resignation. Filmmaker and chess enthusiast Stanley Kubrick loved the game because “it teaches you to think before you act, and to think with the same objectivity when you are in trouble.” Consulting artificial intelligence, a chess engine, of course precludes any thinking at all. It therefore destroys the learning.
But “performance enhancement” is not nearly so clear-cut of a destroyer of learning in resistance sports. Bar all forms of synthetic help, demand that athletes learn with naked brains and “pure” bodies, and you run straight into trouble. Float tanks. Strobe goggles. Wearable sensors. Intravenous hydration. Electronic bike gears. Laser range finders. Digitized film to enhance pattern recognition. Why is pharmacology a more artificial advantage than other technologies employed by wealthy athletes seeking to sharpen their receptors and find that mysterious brain-body synaptic intersection we call “optimal?”
Performance enhancement in sports is not a single problem with a single answer and a single cheating mode but rather a vast, multifaceted series of perplexities. There are different magnitudes of artificialities and offenses, some of which may not be real offenses at all, with varying implications in health, ethics and science. What about the athlete who is simply trying to manage pain. speed recovery or put on lean muscle to deal better with extreme demands? Is it so ethically wrong to minimize self-harm? Does total prohibition needlessly criminalize people who may have a core competitive integrity?
The posing of these questions is not meant to rationalize rule-breaking. Rather, it’s to suggest the subject of performance enhancement could use a hard rethink. To simply categorize it as “cheating” — no different from being told by a computer where to move a knight — out of an anxiety bordering on the superstitious over what is a “natural” vs. “unnatural” body destroys learning, too.
Source: WP
