"When the environment is stable, AI can surpass humans. If the future is like the past, large amounts of data are useful. However, if surprises happen, big data - which are always data from the past - may mislead us about the future. Big data algorithms missed the financial crisis of 2008 and in 2016 predicted Hillary Clinton's victory by a large margin.
In fact, many problems we face are not well-defined games but situations in which uncertainty abounds: finding true love, predicting who will commit a crime and reacting in unforeseen emergency situations are examples. Here, more computing power and bigger data are of limited help. Humans are the key source of uncertainty. Imagine how much more difficult chess would be if the king could violate the rules at a whim and the queen could stomp off the board in protest after setting the rooks on fire. With people involved, trust in complex algorithms can lead to illusions of certainty that become a recipe for disaster."
Trecho retirado de "How to Stay Smart in a Smart World: Why Human Intelligence Still Beats Algorithms" de Gerd Gigerenzer
Sem comentários:
Enviar um comentário