As AI eats the web, the internet’s collective memory is disappearing (thewalrus.ca)

Article summary

The article discusses how the internet's collective memory is disappearing due to the increasing use of AI in search engines, which can lead to inaccurate information and the loss of original content. Google's new AI summaries have been found to invent times, and the company's search results are no longer reliable. The article also highlights the issue of link rot, where pages are erased every day, and the importance of preserving cultural sovereignty. The internet's archival function is breaking down, making it difficult to access accurate information.

Main themes

• AI and search engines
• Internet collective memory
• Information accuracy
• Cultural sovereignty
• Digital preservation

What commenters say

• The decline of Google's search quality is a result of its attempt to be more 'helpful' by serving synonyms and correcting spelling, rather than providing exact results.
• The use of AI in search engines is leading to a loss of original content and the replacement of human-generated information with AI-generated content.
• Some users find alternative search engines, such as DuckDuckGo, to be more effective and less prone to AI-generated content, while others prefer Google's results.
• The reliance on AI-generated content and the lack of human oversight can lead to inaccurate information and a decline in the overall quality of search results.
• The incentive to publish new content is diminishing as AI training bots become the primary visitors to websites, potentially leading to a decrease in the creation of new information.
• The use of AI in search engines can be beneficial for certain tasks, such as configuring a router, but can also hinder learning and understanding by providing answers without context.
• Some argue that the lack of advertising on certain AI-powered platforms is not a long-term feasible model, while others believe that consumers are willing to pay for subscription-based services.
• The future of information retrieval may involve AI companies developing their own training data, potentially leading to a lock-down of information behind paywalls and chatbots.