Jul 23, 2017 · 1 min read
This is the big question.
China can aggressively QA their training data. The rest of the world, not so much. A lot of the data that the “rest of the world” network sees is deliberate errors introduced by everyone from small-timers gaming Yelp and Amazon reviews to state-sponsored misinformation.
This makes it interesting to watch adfraud. If fraud hackers can continue to beat the AI leaders at their core business, then size of network might not be as big an advantage.
