(With the citation and this paragraph in brackets, this essay is more than 600 words but without them its less than 600.)

AI Offers Only Predictions, Not Explanations

	One of the most cited scholars in history, Noam Chomsky, argued strongly against some current trends in the field of Artificial Intelligence. While acknowledging its engineering success, Chomsky thinks that AI direct[s] people away from the original questions. (Chomsky, 2012) Despite the fact that AI predicts whats gonna happen outside the window next much better than the physics department, Chomsky argues, they do not add understanding to the physical world like physics does. Nevertheless, understanding and explaining how the world works has always been arguably the most fundamental motivation for scientific research. If one agrees with this argument, one will also see the essential ineffectiveness of AI solutions to fields like public policy mentioned in Stanfords study, Artificial Intelligence and Life in 2030. 

	For example, the study mentioned CompStat, a system that uses machine learning to predict when and where crimes will happen and who will commit them for the New York Police Department. Since AI now defeats human in the most complex board game ever created (Go), we can reasonably assume that CompStat does its job with very high accuracy. Also assume that it also gives the predictions without any racial or other kinds of biases, a worry mentioned in the study. Perhaps for a month New York City becomes a much safer place: its criminal rate at record low, percent of wanted criminal that are caught at record high, and much more victims get the justice they deserved since AI system in cameras and online collects more evidences to reveal the crime than before. However, if the original cause of crime in New York is not solved, public safety in New York City will return to its original state. For example, if the income gap is still wide, the public education in certain communities is still poor and as a result, the community atmosphere is still bad, those communities will challenge the law again, with ways circumventing AI, probably known as AI itself. Maybe intelligent criminals will invent cop radars that uses machine learning to accurately route themselves for getting around patrolling forces. It takes one to know one, and crimes might only disappear when no one have a reason to commit them. Eliminating crimes from the origin is a never-ending process as well but by treating the cause, rather than outcomes, we are making solid progresses in the long run. This, however, requires negotiation between people across social-economic classes, races and ethnicities. It is the art of compromises and a social and cultural process that involves public discourse and marches. To understand politics, AIs must understand at least human emotions, incentives, and languages. Currently, AIs understand none of these topics. Nonetheless, AI can facilitate this process by means like identifying preferences for every individual by their social media/shopping/search records online and provide useful data that might not be available otherwise. 

	In conclusion, for problems that demands explanations and understandings like science and politics rather than predictions and numbers (problems like transportation and the game of Go), AI can only solve specific cases and alleviate the problem for a brief period. Since AI works around such problems by statistical methods and evolutionary algorithms rather than intuitive reasoning, in the long term anyone that treats AI as a panacea to all such problems will find themselves in a never-ending loop of creating new solutions to what is essentially the same problem. AI so far only offers predictions, not explanations.

Work Cited

Chomsky, N. (2012, November 1). Noam Chomsky on Where Artificial Intelligence Went Wrong [Interview by Y. Katz]. Retrieved September 11, 2017, from https://www.theatlantic.com/technology/archive/2012/11/noam-chomsky-on-where-artificial-intelligence-went-wrong/261637/?single_page=true
(Rep.). (2016, September). Retrieved September 11, 2017, from Stanford University website: https://ai100.stanford.edu/sites/default/files/ai100report10032016fnl_singles.pdf	

