We’ve all likely been in Lee Se-dol’s shoes before, though not on an international stage. After five and a half hours of grueling decision-making and indescribable emotion playing the ancient board game Go, he put his hand against his forehead and announced his resignation from the fifth and final match against Google DeepThink’s AlphaGo AI.
Still running off adrenaline and trying to process exactly what had happened, he seemed almost calm. However, as he studied the middle of the board, playing back previous moves in his mind, anguish set in. He had lost, presumably, way back then, dozens of moves ago. He never had a chance past that point. The palm of his hand came up again, this time in a mock fist, and he shut his eyes in a moment of pure despair. It was if something had been taken from him that he would never get back.
While describing this complex cocktail of emotions and relived memories would be impossible to sum up in one word, “emptiness” comes to mind. Before the matches had begun, Lee was energetic, cocksure about accepting AlphaGo’s challenge. “It didn’t take long for me to make the decision because I was so curious about AlphaGo. I only thought about five minutes.” Now, all that surety was gone, and the uncertainty he had begun to feel since the first match became certain: the AI had outmatched him.
Call it hubris, naivety, denial, what have you, the fact is that Se-dol had predicted the wrong outcome, despite doing everything in his power to assure it. Seeing as Google intends on chalking up this victory to a new AI benchmark, all eyes should be on the future for when AlphaGo’s deep neuron AI system begins to have far-reaching consequences for fields like search engine optimization.
Wait, What Does AlphaGo Have to Do with SEO?
If you are not one of our avid blog readers, you may not have had the chance to read our post on how machine learning and AI already has a hand in the search results that appear on Google. Named RankBrain, helps Google’s other algorithms determine which search results should show up based on the query.
Right now, RankBrain’s influence is severely handicapped compared to the human-coded algorithms. Its input only shows up when the other coding is stumped, typically for never-before-seen queries that have unusual subject matter or word combinations. Presumably, RankBrain will gain higher influence and priority as Google feels progressively more satisfied with its abilities.
RankBrain’s goal in many ways is to help. After all, Google has been relentlessly pursuing smarter searching since first opening their doors. Yet, in other ways, RankBrain pits itself against individuals who have want to rank number one no matter what. Like Lee Se-dol, these individuals have a strong motivation to succeed. In many ways, that drive stems from needing to earn one’s keep by propelling their employer to a distinct advantage. In deeper ways, that drive has to do with ego. Though we may not be aware of it, we have just as much of our perception of self-worth wrapped up in our jobs as anything else. When we let ourselves down, the defeat feels crushing.
Most people can pick themselves up from such a loss. They have the luxury of knowing what they did wrong. But Lee doesn’t. He will likely be going over the defeat millions of times in his head, trying to learn from a strategy that by human standards makes no sense. As President of the American Go Association Andy Okun puts it, “AlphaGo is a creation of humans, but the way it plays Go is not.” Reflecting on the loss, Okun also said “it was a shock. Not a surprise, because I think we had no way to predict the outcome. It just felt bad.”
But why? Why did we feel so personally invested in Lee Se-dol’s performance? “I felt emotional and dizzy, and stepped outside for a minute,” U.S. amateur Go player Ben Lockhart told NPR. This reaction even manifested itself on a microscale as AlphaGo made plays that defied conventional logic. The Korea Herald writes that a move in the second game where AlphaGo placed a stone in an area of the board in which little action was taking place, “made human champion Lee leave his seat to take a break, or likely to pull himself together.”
Flustered, battered, scared, perhaps this is the moment Lee truly lost because he had lost the control he thought he had. Simple truths he had taken for granted where dashed inside the millions of “neural” algorithms DeepThink had plugged into AlphaGo.
Garry Kasparov had a similar reaction in 1997, albeit perhaps as the result of a misunderstanding. DeepBlue, Kasparov’s IBM-built AI opponent, made a bizarre move late-game that some now chalk up to a glitch. Kasparov didn’t see it that way. “He interpreted it as a sign of an intelligence superior to his own, and became anxious and unnerved,” according to Mashable.
The Only Thing We Have to Fear?
Contrary to what science fiction has us obsessing over, AIs aren’t likely to feel motivated to conquer us given the current engineering trajectory. Yet, like facing an opponent in a board game, their goals may run contrary to our own. They also represent an opponent that has the distinct possibility of shaking us to our core and rattling loose truisms we had once accepted as truths.
“There’s a disorienting, airless vibe to facing an artificial challenger,” notes Ken Jennings, former Jeopardy! Champion and loser to IBM’s Watson AI. “You feel unexpectedly alone in the spotlight, but at the same time you’re hyperaware of the millions of tech dollars and labs full of anonymous nerds arrayed against you. Your new opponent, unlike everyone you’ve ever played in the past, can never become overconfident or intimidated. There’s no way to play it psychologically at all, because it has no psychology.”
To extrapolate this idea, making the mistake of approaching an AI on your own terms means having to quickly realize you brought the wrong rule book. As RankBrain advances or something replaces it, gaming the search engine system based on your knowledge of current algorithms or even human psychology becomes a losing prospect. AlphaGo’s victory is an unignorable signal that big changes are on the horizon, in the digital marketing realm as well as in society at large. The best we can hope to do is roll with the punches and learn as we go. Or, like Ken Jennings, we can simply accept our fate.
Let EverSpark Be Your John Connor and Fight Back Against the AI
In the midst of all the philosophical musings and kneejerk paranoia that arose out of the match, the one absolute truth in this event is that humanity was still the winner, even if Se-dol wasn’t. Google’s AI human-programmed accomplished what it set out to do, and though we may not understand exactly how it did it, an optimistic precedent has been set for the future.
This ability to persevere shone through in Lee’s triumphant win on the fourth match. While EverSpark Interactive is hardly on the competitive level of an international Go champion, we do have the staying power to learn from any mistakes we make and boost our clients’ visibility on search engines and more broadly online.
For now, these tactics are amazing at getting real digital marketing results. But in the future, as AIs like RankBrain alter the landscape, we are willing to study as much as we can about our opponents and try to keep you on top. And since “come with me if you want to live” isn’t the best pitch, we’ll advocate that you learn more about our digital marketing services so we can fight the good fight together.