Shulou(Shulou.com)06/03 Report--

At the time of the death of this wallet, most people must have stopped releasing themselves for a few days except for express delivery.

So we want to talk about some relatively serious topics, such as the sudden death of a small seller on the eve of Singles' Day.

Shen Wenjiao, an outstanding Chinese designer who has won 11 red dot design awards in six years, fell at his desk on November 10.

Of course, the news didn't even make it to the list of hot searches, and only a few people complained for him. Because he could have achieved success only by virtue of an award-winning Nude coat hanger, but after countless times of plagiarism and imitation, his rights protection was fruitless for two years, his shop was deserted, and shanzhai products became popular all over the network, and finally ended with the factory closing down in arrears of wages.

Shen Wenjiao's article "the original is dead" removes the dilemma of Chinese original design. But two years later, the speaker is like a flash in the pan, while the infringer is still full of holes. No wonder Big V says bluntly on social networks that China doesn't deserve to have originality.

Of course, such a "map cannon" also makes the general public helpless, both theoretically and morally, the public and various platforms support and encourage originators. However, the various obstacles that need to be overcome in the stage of social progress are always getting rid of the bodies of those who struggle.

Is there any way to put armor on them? it may be difficult for an article to find the ultimate answer, but we think it is time to ask this question and try to reach it.

Saving the original, AI said he was "going south to Canada".

Friends who are familiar with us may wonder if we are going to sacrifice the AI again. Indeed, in the past year we have heard of exploding "fake" apps such as AI face changing, AI voice changing, and AI writing essays. Can't it help the original?

Today we sing a "different tune" to talk about the constraints of AI in protecting originality.

One of the difficulties is that it is not easy to collect evidence originally. Today, whether it is a government-managed copyright center or some third-party copyright websites, they can apply for a "copyright certificate" for their own creation as long as they spend a certain amount of money. However, even if there is good evidence, it is still difficult to collect pirated data.

They have been photographed in various exhibitions, stolen pictures on the Internet, and some even take the lead in publishing original plagiarism after watching the creator's live broadcast. More generally, many designers do not even know that their products have been "copied". For example, the cultural creation products of the Forbidden City have been imitated in the crowdfunding stage and are sold online first.

Can machine learning solve this problem? The answer is: yes, but not necessary.

Although the image recognition algorithm can quickly find high-speed similar pictures in the vast ocean of network data, there is a complex judgment process of legal use and plagiarism, and the cost of network-wide comparison for each work is too high. It not only costs a lot of server costs, but also is not realistic in terms of efficiency.

Therefore, it is impossible to expect the AI system to provide real-time early warning, and it seems that the status quo of "original non-reporting, official non-interference" can only last for a while.

AI saves originality, the second difficulty is to judge plagiarism.

Copyright certificates can limit some direct plagiarism, but in most cases, originators will face a more terrible "magic change" than direct plagiarism.

In the scope of protection of existing intellectual property laws, it is often "only to protect expression, not to protect ideas". For example, Chanel was the first to design and lead the classic small black dress, but it can not stop other clothing brands from making small black dresses, it can only restrict each other from making a certain style of small black skirts. By the same token, many creations that cannot be seen can not be effectively protected.

Therefore, although the depth neural network can judge the similarity of the style between the two works through the migration algorithm, in many cases, even if the style of the two works are exactly the same, the content of the painting is original and does not involve plagiarism.

(Nude clothes hanger and its plagiarists)

Even for original works, it is difficult to quantify because of this "qualitative standard". Infringers can easily change a few details to become their own originality. Although machine learning can quickly identify the slight difference between the two works compared with the naked eye, it is extremely difficult to protect the rights of the "magic change".

The nude clothes hanger designed by Shen Wenjiao mentioned earlier has appeared countless "magic modified" versions on some websites, and it is difficult to remove all the shanzhai from the shelves even in the official background. In the famous novel "the Ghost blowing the Lantern," the original author, the world basher, also sued the production team of "the Nine-story Monster Pagoda" of his work. It took four years to rule that the other party tampered with the original work, but the amount of compensation was only 50,000 yuan.

The input-output ratio of safeguarding rights is low, and many original creators can only choose to "hang people" and "plagiarize" on social networks, and rely on moral condemnation to express their unwillingness.

Original protection, what other technological roads can be taken?

We have to admit that although the technical tool of artificial intelligence plays a wonderful role in many fields, it is still very immature when it comes to plagiarizing piracy.

It sounds a little sad, can copyright protection really only rely on self-consciousness, there is no effective way?

To some extent, this examination question is like a comprehensive problem at the end of modern social civilization, which can not be solved by a single technical formula. Nevertheless, it is imperative to introduce new technologies. However, during this period, how to build a "protective shield" around original protection is a kind of art of scientific and technological civilization.

First of all, the new technology forms a resultant force in the form of cross-linking. At present, there are some copyright platforms based on blockchain technology on the market, bringing new hope to original protection. The IPTM time sign above can quickly make a judgment when registering and identifying infringement, and it can be tested when it is on the chain. Ten thousand pictures can generate test results in 2 hours at the earliest, which solves the embarrassment of "lost sheep to mend" in the past.

However, for the registration of original works, it is often necessary to collectively link electronic vouchers with author information, original content and other data, resulting in scale calculation and cost pressure, which is the main reason to dissuade policy makers and commercial forces. In the future, with the transfer and distribution of distributed idle computing power by deep learning algorithm, under the premise of the surge of digital works, further ensure the low-cost and efficient operation of the copyright platform, so that the risk of infringers continues to rise.

Secondly, technology has become an efficient auxiliary tool to assist intellectual property cases to enter the judicial process.

Although screening "ideas" is not AI's strong point, it is good at promoting judicial efficiency, which can directly solve the core pain points of original rights protection-difficult identification, long trial, and time-consuming.

At present, intelligent justice can complete the comprehensive acquisition, in-depth analysis and automatic comparison of material characteristics and other information through machine assistance, and after these procedural and basic tasks are solved, judicial personnel can be freed from the environment of excessive saturation and lack of skills, and focus their main energy on complex links such as hearing and listening to opinions, so as to shorten the case cycle and reduce the cost of safeguarding rights.

Of course, the most fundamental thing is to enable the originators from the source.

Machine learning is unable to perceive "thoughts", and even imitates human works, both pictures and words are very elementary. But at the same time, it can also help creators maximize the value and charm of their ideas.

For example, a series of cool techs created by "essential artists" adobe can automatically help painters deal with all kinds of lighting and matting problems. Building 3D models in FE can complete rendering and other operations in one minute. The less time you spend on these basic functions in the future, the less time you spend on these basic functions, so as to avoid human designers from "overworking to death". The era of competing individual ideas and creative wisdom will come.

Today, with the gradual landing of new technologies such as AI, cloud, blockchain and so on, people's lives are gradually moving towards the problem of intelligence. Machine, society and human nature are forming the new "iron triangle" of future civilization. This may be a new paradox. Only the evolution and blessing of machine civilization can make the wisdom and brilliance of human nature look so different and worthy of praise.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.