Connecting The Dots: AI Is Eating The Web That Enabled It

By: Tom Wheeler (Brookings Tech Tank)

Large language models (LLMs) of generative AI, which were trained using data scraped from websites, are now using that data to reduce the need to visit those same websites. Digital commentator Casey Newton remarked that “the web is entering a state of managed decline,” while The Washington Post warned, “Web publishers brace for carnage as Google adds AI answers.”

From Decentralized Information to Centralized Conclusions

The World Wide Web, created by Sir Tim Berners-Lee in 1989, transformed the internet into a user-friendly network of diverse information sources. Berners-Lee noted that the first decade of the web was marked by decentralization and a wide array of content options. However, over the years, this vision has been challenged by centralized platforms like Facebook and Google that direct user traffic. Now, generative AI is further eroding this vision by reducing web traffic through its ability to generate direct answers from aggregated data.

Generative AI models, trained on the web’s vast information, are now impacting the very websites they learned from. Instead of users exploring various sources, AI delivers conclusions directly, bypassing the need to visit individual sites.

Apple’s recent announcement to integrate OpenAI into Siri, enabling it to provide ChatGPT-like answers, exemplifies this shift. This move positions Apple as an AI-based intermediary, potentially reducing the need for both website visits and even Google searches, for which Apple pays $20 billion annually.

Studies by The Atlantic, University of Toronto, and Gartner suggest that the decline in website traffic highlighted by Pew research could be just the beginning. Generative AI’s ability to provide direct conclusions threatens the fundamental purpose of websites, especially those reliant on commercial support…

CONTINUE READING…