Will AI Destroy the World Wide Web?
2 days ago
- #World Wide Web
- #Generative AI
- #Model Collapse
- The World Wide Web (Web) was invented by Tim Berners-Lee at CERN in 1989 and gained popularity in the mid-1990s with the release of the Mosaic browser and commercial internet availability.
- The release of ChatGPT in 2022 marked another revolution, making high-quality natural-language text generation widely available.
- Google's business model relies on advertising revenue from Web pages, but GenAI threatens this by providing direct answers without needing to visit Web pages.
- If users stop visiting Web pages, advertisers may stop paying for ads, destabilizing the Web's business model.
- Public Web pages are a major source of data for training large language models (LLMs), and their decline could make LLM training more difficult.
- People are increasingly using GenAI to create content, which could lead to LLMs being trained on AI-generated text, potentially causing irreversible defects in models (model collapse).
- The quality of GenAI-generated answers is leading to concerns about the Web's reliability, with some arguing the internet is 'reverting to beta.'
- ChatGPT acknowledges that AI could erode the Web's value unless regulated and used responsibly.