Summary of the Core Content in Plain Language
Recently, Reuters announced its integration with the cloud data platform Snowflake, allowing all companies to directly access their entire news archive from 1987 to the present, along with real-time updates, and integrate them into their own AI systems. This development goes against common assumptions: many believed that AI could already write news, generate summaries, and revise copywriting, making traditional media content less valuable. However, there is now a trend where companies are willing to pay for authoritative news data. The essence of this shift is that generative AI has moved beyond the initial phase of using vast amounts of historical information to become smarter and is now ready to be applied in real-world scenarios. There is a growing demand for real-time, credible factual data, which is shifting the value of the content industry from content processing to data collection, even giving rise to a new market dedicated to serving AI.
---
Detailed Explanation of Each Point
1. The First Counterintuitive Point: If AI Can Write News, Why Do Companies Still Buy Reuters’ News Database?
Many people are confused by this. The reason is that AI requires two different types of data:
- The first type is like “textbooks for learning”: this includes web pages, books, old news, and papers, which are used to teach AI language logic, basic knowledge, and reasoning methods, essentially training AI to be like a knowledgeable individual with a university education. For example, AI can understand what raising interest rates means and how it affects the stock market based on this historical data.
- The second type is like “new exam questions”: these are real-world changes that occur as AI is used in practical applications, such as recently released financial reports of companies, flight delays, or the last remaining item of a product being purchased by someone else. These events are not part of the AI’s training data, and even the largest models cannot predict them based on historical knowledge.
Previously, AI was mainly used for chatting and answering basic questions, and the lack of the second type of data was not a big issue. But now, as AI is used for tasks like investment research, booking flights, shopping, and managing inventory, relying on outdated information can cause problems. Therefore, the need for real-time, authoritative news from sources like Reuters becomes essential.
2. AI Has Reduced the Cost of Content Processing, Making “New Facts” More Valuable
AI has significantly reduced the cost of tasks that involve processing existing information, such as writing summaries, translating, and rewriting text. A study by Science magazine showed that highly educated professionals could save 40% of time and improve the quality of content by 18% when using ChatGPT for medium-complexity tasks. However, AI cannot attend company meetings, monitor market prices in real-time, or interview reporters about recent events. These tasks of collecting new facts from scratch remain costly; companies need to hire reporters globally, build real-time data collection systems, and verify information to avoid errors. AI can cut and format content efficiently but cannot generate fresh data on its own.
3. The Shift from Selling Content to Selling Data: A New Billion-Dollar Business
Previous collaborations between media and AI companies focused on buying historical data, such as the $250 million five-year agreement between OpenAI and News Corp. in 2024. Reuters, on the other hand, is selling real-time data streams. AI can use this data immediately without waiting for retraining, and companies pay for the speed of data updates, the stability of the interfaces, and the ability to integrate it directly into their systems. This represents a significant shift in the business model.
4. A Major Reorganization of the Content Industry
This new business model will lead to a clear division of roles within the content industry. Companies that can collect and verify facts will become more valuable, as they provide irreplaceable information. Those that only process existing data will become less important. The gap between these two types of companies will widen, as the value of factual information increases and the need for AI to process it decreases.
5. The New Competition in the Age of Intelligent Agents
The competition in the AI industry will not be about who is the most intelligent but about who can obtain the most accurate real-time data quickly. Companies that can set up infrastructure to connect with the real world will gain a significant advantage, as they will be able to provide AI with the latest information needed to perform tasks efficiently. This will determine who benefits from the next wave of AI applications.