No privacy era
Internet society, individuals have no privacy at all, reselling information and telecommunications fraud abound, but this is only the tip of the iceberg. Data is gold, manufacturers spend high cost to collect user data are used to do what?
Although letting others know their own house number and telephone number is a terrible thing. However, it is not always clear that information is most valuable, because it is likely that the user is deliberately forging or inaccurate: you are likely to tell the wrong way to contact the real estate agency; instead, to be alert, implicit and natural, to portray the user’s profile : Weekly browsing the news, brushing social media, cycling and payment sharing mode, all contain a wealth of information. Data generated by user behavior is like a 3D printer, shaping a full-bodied person.
In order to obtain the hidden behavior, different applications in as many as possible in almost all corners of the user behavior as detailed as possible in addition to location, read text messages and phone calls into the application every moment, it will faithfully record each reading , Click and stay, more rogue applications will also record the activities of other applications. All kinds of so-called security assistants hide the background, sucking greedily and transmitting data.
We may not understand ourselves, it is more difficult to get acquainted with others, but the machine knows everyone. Artificial Intelligence is so popular that its biggest application is not yet as cool as science-fiction movies. By mastering the detailed behavior of hundreds of millions of users, complex depth models with hundreds of millions of parameters can learn similar user characteristics and behavioral patterns for Business Marketing and Forecasting. For example, when we see the word “殚 精”, our mind naturally pops up. If user behavior is coded as characters, the machine can estimate the later actions with a high probability after analyzing enough sequences of actions. The hacking into the application transfer behavior and the average person is completely different, he will first check the balance, turn over a list of friends, and then see the binding of which credit card, AI background has already started early warning mechanism, and thus require authentication. These technologies are widely used in anti-fraud cash surpluses and so on.
Under the same sky, everyone’s interests and hobbies are very different, that is, the so-called Thousands of Noodles. User information is heavily used for personal recommendation and behavior prediction. But these technologies will be deeper than expected to all aspects of life.
An e-commerce APP from the installation start, home page layout, ad sorting, to the details page layout, and even logistics are customized. For example, if the regular return, return insurance because they do not make money will not appear in the payment page. Because users with similar personalities probably have similar product preferences, through precise collaborative filtering, precise marketing can greatly improve the click-through rate and purchase rate of advertisements and help manufacturers to obtain more profits.
Information APP Pursuit of clicks and ad exposure. The more eye-catching the “Shocking” title party, the more favored the algorithm, resulting in non-nutritive segments running rampant; the more clicks on a certain type of article, there is a higher probability of recommendation, the strong Constant Matthew effect is serious . For example, engaged in artificial intelligence, then the world is full of artificial intelligence how the article will naturally be formed may be false prosperity and illusion. Parents sharing a daily generation of health chicken soup and rumors, that fundamental reason is that they like it? Or because the information is flooded around the recommendation, so that they have no choice?
Because the current personalized recommendation algorithm is not perfect, the model is very difficult to model the user satisfaction, often there will be users do not like or outdated recommendation. Compared to other applications, WeChat has done a very restrained marketing recommendation. Any big V does not have the privilege of occupying the screen of others. However, we filtered out and filtered similar recommendation systems when filtering friends. Therefore, people of two strata are likely to be in different worlds and do not understand each other.
More than ten years ago, we got information through TV broadcasts and bought goods at supermarkets, all of whom we heard and heard the same. But now everyone looks down at the phone, but the phone is very, very private, you only know your screen, we really are the same?
Is the price of the product really the same? Although the taxi application price, because of the surrounding supply and demand, time and place to dynamically adjust, but it is hard to say whether it is also personalized by personal custom. The tolerance of high prices and the repeated opening of taxi pages will demonstrate the strong desire to take a taxi and further enhance the possibility of fare increases. Taxi around obviously, but the software is not displayed, send a single or probably their own vehicles, because manufacturers do not make money in the taxi.
Marketing is the lifeblood of business, credit is the core of finance. Everyone in the online loan interest rates and quotas is the same? Education, vehicles and real estate will be used to assess credit ratings. For individual micro-loans, overdue users generally have repayment ability but lack of willingness to repay. This situation is difficult to cover the traditional credit, but it is easy to speculate through the usual pattern of behavior. If we do not talk about credit on weekdays and there is no plan for consumption, the algorithm model can already be classified as a high-risk group for a long time. Borrowing quotas will be lower and interest will be higher.
The algorithm of maximizing profit develops the price discrimination of “student card half-price” to incomparable precision. It accurately estimates the highest threshold of user response. We tried to open the hole, if the software deliberately temporarily modified the volatile commodity prices, we have the probability of later found this problem?
A product tips out of stock, it is really out of stock status? Perhaps the system estimates that this one is not profitable, or that buyers are wiping wool. Red envelopes reward seems random, but the lottery on the screen are basically lies. Lucky Goddess will always make the future more profitable people winning. Due to the degree of fatigue control, once won the prize, then pumping hand cramps is also difficult to repeat the lucky.
An article has been deleted, it is really deleted? Or only a few people can not see? An article that pushes a success Nobody appreciates praise, but is it really pushed to someone else? The major media companies have behind the gold and investors, such as in the know is difficult to see the true meaning of Tencent negative information, but Ali Department of the various black materials are rapidly ignorant in the knowledge, you guess Who knows what investment? interesting!
Watch out for control
Due to the current high cost of data acquisition, low back-flow rate, and the technical level of algorithm engineers, the existing algorithm system is far from adequate for controlling and understanding information. In the near future, however, sensors and the Internet of Things (IoT) in all corners collect vast amounts of information; real-time computing becomes mainstream, and the model reacts to new signals within seconds to adjust tactics; GPU clusters of tens of thousands of machines Be able to easily simulate the mind and behavior of hundreds of millions of users. Maybe just a sudden drifting brain needs, precision ads will be immediately displayed to you. You must see what you or others want to see. Everyone is covered in a shell around the periphery, to see the world is processed filter, the information becomes unequal, the world is further divided.
Three primary colors incident was indignant geothermal for a week, but then completely disappeared in the network, people have to confuse the power of information control. Currently sensitive information monitoring is still stuck in very original keyword filtering, it is easy to avoid shielding by replacing sensitive words, let alone sensitive audio and video. However, future algorithms will have the ability to understand beyond the average person. If a piece of information is made inappropriate, then it will be blocked from the source, whether it is the so-called paragraph, metaphor or irony, explosive communication is nonsense.
In the second quarter of the British drama “black mirror”, a few stories are told. One of them is that in the future, everyone in the play is fitted with a smart eye. As long as you do not like a person, you can shield him directly. Later, he will The outline of the figure that turned pale before you, though he could hear his voice, did not know what he said specifically. The divorcee divorced his ex-wife and the ex-wife hated him, so his ex-wife shielded all the information about her and her daughter. Unloved man wanted to see her daughter side, begging, appears in the eyes of her daughter is a strange white shadow and messy noise.
Perhaps we can systematically read information and knowledge to avoid fragmented custom information on the brain; maybe we can try more face-to-face temperature conversations rather than hasty words on the screen; maybe we can spend more time on our life And travel instead of gossiping someone else’s story in the drama movie.
We are still glad that the vast majority of recommendations and customizations have not risen to the level of “control,” so there is always free access to information through other means; luckily the controller is a human rather than a more powerful machine, or the story in the Matrix Will be a reality; fortunately, their motives are generally out of economic interests rather than political purposes, or wherever they are peace and prosperity and peace and prosperity.