Written as part of the MATS 9.1 extension program, mentored by Richard Ngo.

From March 9th to 15th 2016, Go players around the world stayed up to watch their game fall to AI. Google DeepMind’s AlphaGo defeated Lee Sedol, commonly understood to be the world’s strongest player at the time, with a convincing 4-1 score.

This event “rocked” the Go world, but its impact on the culture was initially unclear. In Chess, for instance, computers have not meaningfully automated away human jobs. Human Chess flourished as a pseudo-Esport in the internet era whereas the yearly Computer Chess Championship is followed concurrently by no more than a few hundred nerds online. It turns out that the game’s cultural and economic value comes not from the abstract beauty of top-end performance, but instead from human drama and engagement. Indeed, Go has appeared to replicate this. A commentary stream might feature a complementary AI evaluation bar to give the viewers context. A Go teacher might include some new intriguing AI variations in their lesson materials. But the cultural practice of Go seemed to remain largely unaffected.

Nascent signs of disharmony in Europe became nevertheless visible in early 2018, when the online European Team Championship’s referee accused a player, Carlo Metta, of illicit AI use during a game. His results were voided and he was banned from further participation in the event. At the time the offending game was played, open-source engines based on the AlphaGo paper, such as Leela Zero, had only been around for about a month. However, a predecessor called Leela 0.11 was already widely available and was known to match the level of the top Europeans that Metta was facing. Metta’s accusers claimed that his play was too similar to this AI’s preferred moves. It was moreover considered suspicious that his Over-The-Board (OTB) play agreed significantly less with the AI than his online moves did.

Unfortunately for the prosecution, their results were reported in intransparent and sloppy ways. This is evidenced by the fact that the best compilation of their findings is the slapdash facebook thread I linked above. This, along with the circumstantial nature of the evidence, was criticised in the same thread by community members. Teammates and friends of Metta’s also stepped up to publicly defend him. One way in which their rhetoric proved effective involved the public...

Wenngle

Wenngle