The author thinks that
Supervise the production
Ordinary people who are trying to fight algorithms understand , It's a battle of great power .
You and I are in the siege
“ You can believe ？ Players who can still rank at this level in the king's rank ！”
Almost every few days , Wheat will be because of 《 Glory of Kings 》 I'm in a hurry . Often single row up, he thinks the strength is enough “ Take flight ”, But the winning rate is always 50% about . Especially in long-running games , There will always be “ Win one, lose two ” The situation of . Maizi says frankly , I'm not the one who can't afford to lose , It's just fantastic to lose so often . such as , After winning two or three games in a row, you'll meet players who don't look at the lineup and grab people , Or it's the low quality spurt that disturbs the mood of teammates . Even if you play well , I can't win .
“ I thought the quality of the game would be higher when I went to the king's rank , I didn't expect it to be like this . It's like there's an invisible hand around the war .”
By this “ The king's sanctions Bureau ” Not a few players are bothered , Some even claim that their teammates or opponents are AI robot , Others say the damage done by themselves in the sanctions Bureau will be significantly reduced , The key moment always can let the opponent silk blood escape .
Put aside these magical ideas, no matter , quite a lot MOBA In the game, there is a set of “Elo Integral algorithm ” It's working . This algorithm is to increase the user's game duration in disguised form by controlling the player's speed , Constantly create “ One God with four pigs ” or “ Four gods protect one pig ” The game environment of . Once the player's game is brilliant , So-called “ Hidden points ” It's going to go up a lot , The probability of matching players with low hidden score is greatly increased . in the course of time ,Elo The mechanism becomes “ Protect the chicken , Crush the strong ” The existence of .
To counter systemic sanctions , Some players have summed up a methodology for deliberately suppressing hidden points . For example, learn to “ Show weakness ”, When you can fish and win, don't carry; don't take MVP Select fewer core output bits ; To match before ranking, deliberately lose a few games, etc .
But you know , One of the most attractive things about competitive games is that it provides “ Win by skill ” The fair stage of . If algorithmic rules deprive this core experience , Even users need to do the opposite , So essentially and “ Playing match fixing ”、“ Fake Boxing ” What's the difference ？ This imperceptible value orientation , Long gone in the right direction .
A few days ago , When wheat was in a bad mood, he played a game . Crispy picked two teammates , Grab online resources like crazy , This leads to frequent killing of Wheat by the opponent . Driven by negative emotions , Mai Zi refused to join the whole sleepwalk . But even if it's such a hindrance , The game still won without suspense .
After the game, wheat quietly unloaded the game . Although he knows , I will download the game again one day .
The unbalanced algorithmic balance
When did the algorithm stand opposite to the user ？
Maybe , In search engines and e-commerce platforms, rules are captured for information and goods through competitive ranking “ Open the back door ” when , The hidden danger has been buried . Because it's equivalent to building a platform “ Fairness ” As the basic principle of trading chips .
With the expansion of platform boundary and volume , This controversial principle has been brought to more scenarios .SEO、SEM It's become a business , Information platform hot search 、 E-commerce platform business rankings can be improved through some “ Operation method ” To get there quickly . E-commerce promotion Festival is no longer “ cheap ” The pronoun of , It's even more impossible to prevent big data from becoming mature . What is more , Many people make complaints about the platform , The key words mentioned in the oral chat between forefoot and friends , The hindfoot appears in the push information of an application .
Many users have just learned “ The inverse algorithm ” Consumption in style . For example, we will shut down our paid member service on a certain platform from time to time , Switch to the competitive platform , In this way, the algorithm is repeatedly obtained as recall “ Lost users ” Targeted subsidies sent out .
For these “ Counter algorithm ” The needs of , There are endless price comparisons 、 The voucher collection tool also came into being , In order to help users against the algorithm to create “ At a low price ” False appearance . The embarrassment of the , This kind of application is essentially a transfer of commercial value , It's not about algorithmic values “ Righting ”： The price comparison software attracts traffic through the price comparison function of the whole network , Then divide it into 、 Intelligent algorithm, advertising and other forms of profit .
There always seems to be one “ have sinister motives ” We can't avoid it . It's not that there are people who have successfully defeated the rules of the algorithm , It's just their “ victory ” But it doesn't make people happy .
This year, 10 month , Consumers rely on illegal software repeatedly swipe App integral , Recharge your iqiyi members to 2111 year . Similarly, , With the help of a programmer's friends, a gamer Yang plays in the game by abnormal means “ Plucking wool ” the height is 400 Ten thousand yuan . in real life , With a large number of phone cards to constantly change people's settings 、 Rely on a variety of snap up plug-in software 、 Get points through virtual transactions 、 Consumer coupons “ Technical cattle ” Not in the minority .
More Than This , Professional bad evaluator 、 The problems of swiping bills and credit speculation , Also rely on algorithm loopholes to earn a lot of improper benefits . No matter how the system algorithm is updated , The relevant grey industry chain can always “ A stirring among the dry bones ”. It's just over GeekPwn 2020 International Security geek competition , People are even more surprised to find that , Many cutting-edge algorithms can be cracked . Let the autonomous vehicle hit the wall actively 、 Embezzlement of funds account 、 Cheat face recognition to get other people's decision-making power . The operation of the mysterious killer organization in these sci-fi movies , It's all coming true .
From this point of view , The algorithm creates another imbalance ：“ professionals ” Like a duck to water , Ordinary people can only bear with it .
More subtle side effects
1999 year , There was a dream Chinese website “72 Hours of Internet survival ” test , Let the tester in a completely closed state , Live on the Internet 72 Hours . At the time , Many testers don't understand online shopping , You can't even get food or quilts , Had to quit the test halfway .17 After year , There's another one “72 No network to survive ” The test is coming on stage . But the test results are quite the opposite ： Participants in the test have said that “ to be in deep anxiety day seems like a year , It's very painful ”.
The growing dependence of the public on smart devices such as mobile phones is doomed , The connection between human and algorithm will not be easily broken . Except for visible problems , Other side effects of the algorithm are also spreading .
The most intuitive problem is data privacy .
The starting point of the Internet age is called “ No one knows if it's a man or a dog behind the computer ”, But now the reverse is true . In a sense , The user in front of the algorithm is equivalent to naked . I remember a case in a related book ： A father complained to an e-commerce platform , Because the platform frequently pushes the subscription information of maternal and child products to her minor daughter . But a few months later, the father called to apologize ： He mistook the platform , My daughter is really pregnant .
Imagine , Telemarketing that pinpoints our consumption or browsing history on the phone , We can even reason out our financial situation and sell related products , This unequal information gap can put ordinary people in a passive situation . God knows that his information assets are tossed and turned between different algorithms , What kind of matching target will it point to .
Concerns about data privacy , It exacerbates people's sense of insecurity .
2019 year , A big data survey released by Beijing Consumer Association shows that ,56.9% Of the respondents said they had been killed by big data ,88.3% People said that big data is common . Maybe this 88.3% You can't tell me “ conclusive evidence ” To prove that you have experienced the phenomenon of big data killing , But the sense of distrust and the sense of confrontation are lingering .
You bet , If you're connected to a cell phone App、 Microphone 、 The camera may be inadvertently “ sell ” own ,“ A sense of security ” How to start ？ Plus countless big data ripening Events , Let people almost cut off the right “ Enterprise self-discipline ” The expectations of the . In the short term , These problems will not hinder the realization of business value of enterprises . But in the long run , If you can't eliminate people's concerns about algorithms “ I can't stand it and I can't do it ” The mood of , The road to popularization of new technologies and algorithms for future scenarios will be more tortuous . At the time , It may not be that you want users to regain confidence in the algorithm “ Red envelope Marketing ” It can be solved .
The bigger problem is , The improper use of algorithms distorts the mutual trust between people 、 mutual understanding 、 The side of helping each other .
For example, in games , A player who makes a mistake will identify his teammate as “ System algorithm ” The tool person who uses it to obstruct oneself from going up . Even if it can turn the war around , We will also be in each other's resentment in the hasty surrender , Hope for the next match algorithm ; In a social environment , Deceptive beauty algorithm 、 The marriage website algorithm that over label people has planted too many seeds of suspicion , And then there's the wrong words of the rumor makers , In the promotion of the algorithm, the innocent will die irretrievably ; In venture capital and financial scenarios , The algorithm will credit 、 Traffic and other resources are over poured into the icing on the cake , But not willing to do nothing “ Excellent record in history ”, Or for those who are struggling with some small blemishes .
In recent years , There are always “ Whether the poor family can give birth to a noble son ” The discussion of the . Just imagine , If we overuse algorithms as gatekeepers and decision makers in key links , So it's essentially a society that refuses to believe in miracles , The morbid reason is not willing to provide a channel for small probability events .
Of course , The social and commercial value of algorithms is obvious to all . We focus on the side effects of algorithms , Not deliberately demonizing the digital age driven by algorithms , It's the hope that people can take the initiative to become regulators of intelligent algorithms , It will be more in line with the spirit of the development of algorithms .
Casey · O'Neill in 《 Algorithmic hegemony 》 We use such a sentence as the opening remarks in Chinese ： This book is dedicated to digital refugees . But in the end , She still thinks that the public can eventually tame algorithms by uniting .
so , Even when it comes to specific problems, it's very difficult to compete with algorithms , But there is still reason to believe that ： Every challenge to the algorithm makes sense , Those little efforts will eventually become a glorious and bloodless revolution . Need to know , Even 《 Biochemical crisis 》 The super algorithm in AI The Red Queen is in the hero's “ Data tuning ” I betrayed the umbrella company , Why should we be too pessimistic ？