![\"Is](\"https:\/\/hackernoon.com\/hn-images\/1*MYyRd91svmNrVUz6uDAq2A.png\")
If your dictionary defines AI<\/a> as magic or robots (or magical robots), of course you\u2019ll be disappointed when it doesn\u2019t deliver the cure to all that ails you. Let\u2019s look at three common gripes using simple examples everyone can grasp.<\/p>\n\n\n\n A respectable software engineer once asked me with a straight face, \u201cCan AI know that Canada is a country?\u201d<\/em><\/p>\n\n\n\n Hold your horses there, cowboy. Let\u2019s take a moment to think about how you<\/em> know that Canada is a country. Someone told you that fact when you were little, you memorized it, and now you\u2019re looking it up.<\/p>\n\n\n\n We can write the code that does that without AI\u2014record the data in a table, then if someone asks about Canada, the program looks the word up and outputs the answer. What do you need AI for here?<\/p>\n\n\n\n Nothing, that\u2019s what.<\/p>\n\n\n\n If you expect that AI<\/a> is magic, you\u2019ll try to use it for everything<\/em>. When your bosses find out how much effort you\u2019ve wasted coming up with a complicated solution to a simple problem<\/a>, it\u2019s hard to blame them for thinking that AI is hype and nonsense.<\/p>\n\n\n\n If you can solve a problem without AI, then don\u2019t use AI. Seriously, no matter what my employer<\/a>\u2019s marketing team says at you.<\/p>\n\n\n\n AI is like medicine\u200a\u2014\u200ait can be a life-changer to those who need it, but everyone else should know better than to snack on it out of boredom.<\/p><\/blockquote>\n\n\n\n Pick your use case<\/a> and think about it carefully<\/a> before<\/em> you collect any data or hire any PhD gurus. If you feel the need to apply AI somewhere\u200a\u2014\u200aanywhere!\u200a\u2014\u200ajust because all your friends are doing it, you\u2019re setting yourself up for failure.<\/p>\n\n\n\n If you can make it work without AI, so much the better.<\/p><\/blockquote>\n\n\n\n Instead, start with a use case<\/a> you care about and if you can make it work without AI, so much the better.<\/p>\n\n\n\n Canada, I\u2019m not done with you yet.<\/p>\n\n\n\n In our quest to determine whether AI can know that Canada is a country, we established that a computer can store and retrieve such information without any fancy-pants AI<\/a>. Our engineer friend wants to kick it up a notch: \u201cCan a machine learn that all by itself?\u201d<\/em><\/p>\n\n\n\n Whoa, what do you mean by \u201clearn<\/em>\u201d and \u201call by itself<\/em>\u201d? Those words mean different things to different people. Let\u2019s answer this version: \u201cCan we expect a machine reliably output the conclusion that <\/em>Canada is a country if it never had access to the word <\/em>Canada before?\u201d<\/em><\/p>\n\n\n\n Hey people-who-can\u2019t-read-Chinese, is \u9999\u8549 a country? How about \u82f1\u56fd? No, don\u2019t go looking up the answer, that\u2019s cheating. You have to learn this all by yourself,<\/em> remember?<\/p>\n\n\n\n When you have no additional information, how could you possibly know the answer? Similarly, common sense should force you to suspect that AI can\u2019t learn things if there\u2019s no information to learn from. You\u2019d be correct.<\/p>\n\n\n\n AI is all about extracting patterns from information and using those patterns to automatically make a recipe for turning your next input (Canada) into an output (country). So let\u2019s ask ourselves: what relevant patterns could our computer possibly use if it has never seen the word before?<\/p>\n\n\n\n If there was nothing to learn from, learning is impossible.<\/p><\/blockquote>\n\n\n\n Even if we have some data, our algorithm<\/a> might pull out spurious patterns<\/a> that give us a stupid recipe. Let\u2019s imagine these are our training data: South Africa-country, hippopotamus-animal, frog-animal, Russian Federation-country, United States-country, cat-animal, United Kingdom-country, raccoon-animal, South Korea-country, New Zealand-country, butterfly-animal, giraffe-animal.<\/em><\/p>\n\n\n\n Before you\u2019ve even finished reading the first pair, your AI algorithm has already digested them. It gives a satisfied burp and invites you to input your noun. Any guesses what it does when you show it Canada<\/em>?<\/p>\n\n\n\n There are two loud patterns in these data. One is that all the countries have capital letters. If that\u2019s the basis for the AI\u2019s recipe, then \u201cCanada\u201d would be labeled correctly, but \u201ccanada\u201d would not.<\/p>\n\n\n\n What if the recipe were based on a different pattern?<\/p>\n\n\n\n Did you notice all the countries have two-word names, while animals are single words? Well, your algorithm did. It says that Canada is clearly an animal. Oh deer.<\/p>\n\n\n\n Yes, machine learning<\/a> doesn\u2019t work\u2026 if there\u2019s nothing to learn from. AI<\/a> is not for people who can\u2019t get their hands on relevant data<\/a>. If you want your solution to work well for all countries, your dataset can\u2019t only have the two-word ones in it.<\/p>\n\n\n\n Garbage in, garbage out.<\/p><\/blockquote>\n\n\n\n It\u2019s important not to lose your grip on common sense here; the basics of learning and teaching that apply to human students also apply to AI. Datasets are textbooks and you\u2019re the teacher.<\/a> If you give your students garbage textbooks, expect them to learn some garbage.<\/p>\n\n\n\n If AI<\/a> isn\u2019t magical robots, what is it? It\u2019s just a tool that helps you write code<\/a> for tasks that you\u2019re struggling to express the instructions for<\/a>.<\/p>\n\n\n\n Why would a good programmer have difficulty with giving a computer instructions? Isn\u2019t that the job? Sure, but some tasks require extremely complicated instructions. If they\u2019re too complicated for the human mind to wrap itself around, you won\u2019t be able to come up with them\u2026 unless you can communicate the task a different way.<\/p>\n\n\n\n AI lets you make your wishes known with examples (data<\/a>) instead of explicit instructions<\/a>. That means you have a shot at automating tasks that you can\u2019t write instructions for.<\/p>\n\n\n\n Simple solutions don\u2019t work for tasks that need complicated solutions. So AI comes to the rescue with\u200a\u2014\u200asurprise!\u200a\u2014\u200acomplicated solutions.<\/p><\/blockquote>\n\n\n\n It also means that you should start expecting a tangle of complicated instructions<\/a> when the AI algorithm distills patterns into code<\/a> for one of those headache tasks. When you read the recipe it came with for you\u2026 it\u2019s unreadable.<\/p>\n\n\n\n Many people have a gut reaction to mystery and ambiguity: \u201cGet rid of it! Simple or I don\u2019t want it! I can\u2019t trust it.\u201d<\/em><\/p>\n\n\n\n Wishing complex things could be simple doesn\u2019t make them so.<\/p><\/blockquote>\n\n\n\n It looks like you\u2019re stuck with two bad options: live your life solving nothing but the simplest problems or progress beyond the low hanging fruit, but give up trust. Luckily, there\u2019s another way.<\/p>\n\n\n\n We need a better basis for trusting those machines than reading unreadable things. The good news is that there is one: testing<\/a>.<\/p>\n\n\n\n You don\u2019t need to understand how<\/em><\/strong><\/a>it works to check that it does<\/em><\/a>work.<\/p><\/blockquote>\n\n\n\n Proper testing<\/a> isn\u2019t trivial<\/a>, but it\u2019s a lot easier than making sense of something so huge it makes you dizzy. It\u2019s also a principle that we practice often\u200a\u2014\u200afor example with medicine<\/a>. Do you know how that headache pill works? Neither does science<\/a>. The reason we trust it is that we carefully check that it does <\/em>work. (Here<\/a>\u2019s my deeper discussion of testing as a basis for trust.)<\/p>\n\n\n\n If you\u2019re content with only solving easy problems, it\u2019s okay to sit this AI thing out.<\/p><\/blockquote>\n\n\n\n The problems of the future will only be getting harder. After you automate the simple tasks, you\u2019ll want to move on to bigger challenges. Once you reach past the low-hanging fruit, you\u2019ll run into a task you can\u2019t solve using your old tricks and the brute force of raw imagination. You\u2019ll realize you can only communicate what you want with examples<\/a>, not instructions\u2026 Welcome to AI.<\/a><\/p>\n","protected":false},"excerpt":{"rendered":" Is AI a fad? AI isn’t a fad, it’s the way to progress. The problems of the future will only be getting harder. After you automate the simple tasks, you’ll want to move on to bigger challenges.<\/p>\n","protected":false},"author":335,"featured_media":9242,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"content-type":"","footnotes":""},"categories":[183],"tags":[226,97],"ppma_author":[2050],"class_list":["post-9241","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-ml","tag-ai","tag-artificial-intelligence"],"authors":[{"term_id":2050,"user_id":335,"is_guest":0,"slug":"cassie-kozyrkov","display_name":"Cassie Kozyrkov","avatar_url":"https:\/\/www.experfy.com\/blog\/wp-content\/uploads\/2020\/04\/medium_df35f80d-2bff-4fe3-b741-a94d51320e00-150x150.jpg","user_url":"https:\/\/careers.google.com\/?src=Online\/LinkedIn\/linkedin_profilepage&utm_source","last_name":"Kozyrkov","first_name":"Cassie","job_title":"","description":"Cassie Kozyrkov is Chief Decision Scientist at Google, Inc. With a unique combination of deep technical expertise, and world-class public-speaking skills, she has provided guidance on more than 100 projects and designed Google's analytics program, personally training over 15000 Googlers in statistics, decision-making, and machine learning.\u00a0"}],"_links":{"self":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/9241","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/users\/335"}],"replies":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/comments?post=9241"}],"version-history":[{"count":3,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/9241\/revisions"}],"predecessor-version":[{"id":10670,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/9241\/revisions\/10670"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media\/9242"}],"wp:attachment":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media?parent=9241"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/categories?post=9241"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/tags?post=9241"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/ppma_author?post=9241"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}Fad gripe #1: \u201cAI is a waste of time\u201d<\/h3>\n\n\n\n
![\"Is](\"https:\/\/hackernoon.com\/hn-images\/0*DbJXPzTQKhKujIpc.jpg\")
How to avoid trap #1<\/h4>\n\n\n\n
![\"Is](\"https:\/\/hackernoon.com\/hn-images\/0*33RB8EMrjYdap9np.jpg\")
Fad gripe #2: \u201cAI doesn\u2019t work\u201d<\/h3>\n\n\n\n
![\"Is](\"https:\/\/hackernoon.com\/hn-images\/0*GNKnKhI6wlVGH2RH.jpg\")
![\"Is](\"https:\/\/hackernoon.com\/hn-images\/1*jV-9Elm1zsTQaHhWC1dMpQ.png\")
![\"Is](\"https:\/\/hackernoon.com\/hn-images\/0*7WD0i-5SFddDRzuH.jpg\")
How to avoid trap #2<\/h4>\n\n\n\n
Fad gripe #3: \u201cAI is untrustworthy\u201d<\/h3>\n\n\n\n
How to avoid trap #3<\/h4>\n\n\n\n
![\"\u2019re](\"https:\/\/hackernoon.com\/hn-images\/0*4M6qp3HGFIosLBgb.jpg\")
<\/figure><\/div>\n\n\n\nAI isn\u2019t a fad, it\u2019s the way to progress<\/h3>\n\n\n\n