テクノロジー
AI学習データの「対価」をどう定義するか：Creative Commonsが提唱する「Pay-to-Crawl」7つの原則と、崩壊するWebの社会契約
2025年12月、インターネットの「オープン性」を象徴する非営利団体 Creative Commons（クリエイティブ・コモンズ、以下CC）が、歴史的な方針転換とも言える声明を発表した。AIによるWebコンテンツのクロー […]
Term
Pay-to-Crawl
別名: ペイ・トゥ・クロール
Overview
Pay-to-Crawl is a technical and economic framework where AI developers and web crawlers pay a fee to access and use website content for training models or other purposes. It aims to restore the balance of the web's social contract, which has been disrupted by generative AI's ability to extract value without returning traffic to original creators. The system is designed to automate licensing for small-to-medium publishers who lack the bargaining power of major media outlets.
Mentioned Articles
1 件External Mentions
5 件- arXiv Dynamic Grammar-Compressed Self-Index in $δ$-Optimal Space
- arXiv Pay-Per-Crawl Pricing for AI: The LM-Tree Agent
- arXiv Enhancing Dialogue Generation via Multi-Level Contrastive Learning
- arXiv Analysis of the Web Graph Aggregated by Host and Pay-Level Domain
- arXiv LaMMos - Latching Mechanism based on Motorized-screw for Reconfigurable Robots and Exoskeleton Suits