AI summarization technology is triggering a transformation in the content ecosystem. It allows users to quickly access information, but it also raises issues such as copyright infringement, traffic loss, content value dilution, and business model impact. This article delves into the impact of AI summarization and where it is headed in the future.
I recently discovered a very interesting phenomenon.
Turn on your phone and swipe the news, do you often see various “AI summaries”? For articles of hundreds or even thousands of words, AI will summarize them clearly for you in two or three sentences.
Save time and effort, and be extremely efficient.
You might think, how good is this! In the era of information explosion, AI helps us filter noise and extract essence, which is simply the savior of “information overload”.
But on the other hand, those publishers and media organizations that have worked hard to produce content have “fried the pan”. They protested angrily and even took the AI company to court.
Why is that, you may ask? Isn’t AI helping them spread content?
B-end product manager’s ability model and learning improvement
The first challenge faced by B-end product managers is how to correctly analyze and diagnose business problems. This is also the most difficult part, product design knowledge is basically not helpful for this part of the work, if you want to do a good job in business analysis and diagnosis, you must have a solid …
View details >
I said, you think of AI as simple. AI summaries are not only bringing convenience to users, but also triggering a war between “stealing” and “grabbing”. It’s not just a technical issue, it’s a “life and death” about copyright, value and business model.
Today, let’s talk about this matter: Why did AI summaries spark collective protests from publishers? What kind of business logic and future trends are hidden behind this?
The “summary” you think may be completely different from what I want to say.
1. AI Summary: Why did it trigger a “collective protest”?
AI summary sounds like a “win-win” good thing: users save time, and AI companies provide services. So why aren’t publishers happy?
Because, in the eyes of publishers, this is not a “win-win” at all, but a “salary draw at the bottom of the pot”.
1. Copyright infringement: My “dish”, you directly “served” it?
This is the core point of contention.
The publisher argued that AI companies used their copyrighted content without authorization when training models. Then, the AI directly generates a summary, which is equivalent to directly “serving” the “dishes” they have worked so hard to make to users without paying any “meals”.
For example, the New York Times issued a “cease and desist” notice to AI startup Perplexity, accusing it of using news content to generate summaries without authorization.
“Instead of scraping data to build foundation models, we index web pages and present factual content as citations to provide answers to users’ questions,” Perplexity told Reuters. ”
There are even authors and publishers who have directly sued Microsoft and OpenAI, accusing them of using millions of copyrighted works to train AI models.
2. Traffic loss: If users don’t click on links, what do I rely on to “live”?
In the past, users found news through search engines and clicked on links to news websites, which led to traffic and advertising revenue.
This is one of the main sources of income for publishers.
But now, AI directly presents the summary on the search results page, and the user leaves after reading the summary without clicking on the original text at all. This directly led to a significant drop in publishers’ website traffic and a sharp drop in advertising revenue.
Some bloggers have complained that Google’s AI Overviews feature has caused their search traffic to drop sharply because the AI directly provides summarizing answers, “trapping” users on the search page.
For example, after the launch of Google’s AI overview feature, the search traffic of gift store Artmall plummeted by 40%. The store owner said: “Users stop clicking on the website after reading the summary, which seriously affects our exposure and revenue. ”
3. The value of content is diluted: How many words is my “hard work” worth?
An in-depth report can be spent weeks or even months interviewing, verifying, and writing.
But AI can generate a summary in seconds.
This makes publishers feel that the value of their hard-earned content has been severely diluted.
When users get used to free, fast AI summaries, will they still be willing to pay for high-quality, original content? This directly impacts the foundation of the content industry.
4. Business model impact: My “job”, you come to “smash”?
The traditional publishing business model is based on content creation, distribution, and advertising/subscription revenue.
The emergence of AI summaries is fundamentally shaking this model.
Publishers believe that AI companies are using their content to build their own business empires, while publishers are making “wedding dresses” for others. This is not just a copyright issue, but also a matter of survival.
So you see, the protest caused by AI summarization is not just a simple technical friction, but also a deep contradiction between content creators, platform parties and users about “value distribution” and “business model”.
This war is far from over.
2. AI companies: “aggressor” or “collaborator”?
How did AI companies respond to publishers’ protests?
They usually give several reasons:
1. The principle of “fair use”
AI companies believe that they use publicly available web content to train their models, which falls under the category of “fair use”.
It’s like humans learning knowledge, reading a lot of books and articles, and then forming their own understanding and output. Instead of copying and pasting the original text directly, they generate new content.
In recent rulings, some U.S. courts have also tended to uphold that AI companies’ use of copyrighted works when training models is “fair use,” as long as they legally acquire the content.
For example, in a 2024 class action lawsuit, authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson accused Anthropic AI of unauthorized use of millions of copyrighted books, including plaintiffs’ works, to train its AI model, Claude, including scanning paper books to input data.
U.S. District Judge William Alsup found that Anthropic’s use of the plaintiff’s books to train the AI model Claude was in accordance with legal principles and dismissed the author’s copyright lawsuit.
The judge said that Anthropic’s digitization of printed books purchased could also be considered fair use.
However, not all of Anthropic’s training book sources are normal purchases, so in the end, part of the case focused on Anthropic’s liability for using pirated works, which is expected to go to court in December.
From the perspective of future trends, it should be reasonable if AI companies can obtain relevant training resources through reasonable and normal channels, whether it is book content or various commercial data.
2. Improve efficiency and optimize experience
AI companies emphasize that their summarization feature is designed to help users access information more efficiently and improve user experience.
In this era of information explosion, users need more convenient ways to digest massive content.
AI summarization meets this need.
3. The possibility of win-win cooperation
Some AI companies have also begun to try to cooperate with publishers to obtain content usage rights through paid licensing.
For example, Amazon and The New York Times have reached an AI content licensing agreement that allows Amazon to use The New York Times content on its AI platform.
This shows that AI companies are not completely rejecting cooperation, and they are also exploring new business models to try to find a balance between technological innovation and content copyright.
3. Where is the way out?
So, where is the way out of this “war”?
I think the following trends may emerge in the future:
1. Improvement of copyright legislation and precedents
With the rapid development of AI technology, existing copyright laws have lagged behind.
Governments and the legal community need to accelerate the pace of developing clearer laws and regulations to define copyright issues in AI training and content generation.
It will take a long process, but the direction is clear.
2. Explore new business models
Publishers should not only stop at the level of protests, but also actively explore business models that coexist with AI.
For example: content licensing: Like the New York Times and Amazon, directly license content to AI companies to obtain revenue. API interface: Provide standardized API interfaces for AI companies to obtain content through paid calls and summarize or generate it. Differentiated content: Focus on producing deep, exclusive, emotional, and thoughtful content that is difficult for AI to replace, enhancing the scarcity and value of content. User Pay: Guide users to pay for high-quality original content and cultivate users’ payment habits.
3. Intervention of technical means
In the future, more technical means may emerge to protect content copyright, such as content traceability technology, watermark technology, etc., to ensure that the source of content can be traced and used can be monitored.
4. Establishment of industry standards
AI companies and publishers need to establish a common industry standard that clarifies the norms for content use, the mechanism for revenue distribution, and the way to resolve disputes.
This requires both sides to sit down and negotiate together, rather than blindly confronting each other.
So you see, the protest caused by this AI summary is not only a conflict between technology and law, but also a reconstruction of the “future content ecology”.
Whoever can find his place in this restructuring will win the future.
4. Write at the end: AI and content, are they “enemies” or “partners”?
Back to the original question: What exactly does AI summaries mean behind publisher protests?
My answer is: this is not only a conflict between technology and copyright, but also a “big test” of the content industry in the AI era.
The emergence of AI is like a “double-edged sword”.
On the one hand, it greatly improves the efficiency of information dissemination and allows users to obtain knowledge more conveniently. This is an inevitable trend of technological progress, and we cannot stop it.
On the other hand, it has also brought an unprecedented impact on the business model and value chain of the traditional content industry.
If content creators do not get the rewards they deserve, then there will be fewer and fewer high-quality original content, and the entire information ecosystem will ultimately be damaged.
Therefore, we cannot simply regard AI as an “enemy” or blindly embrace it.
What we need to think about is: How to redefine “content value”? In an era where AI can mass-produce content, the uniqueness, depth, emotion, and perspectives of human creation will become even more scarce and precious. How to build a “symbiotic relationship”? AI companies and content publishers are not zero-sum games, but partners who can empower each other. AI can help with content distribution, personalized recommendations, and content can provide high-quality training data for AI. How to improve the “rule system”? Laws, ethics, and industry standards all need to keep up with the pace of AI development and escort the healthy development of AI and content.
Finally, I would like to leave you with a question: In the era of AI, how do you think we should better protect the value of original content while embracing the convenience brought by AI?
