An unnamed large AI company has struck a deal with Reddit to use user-generated content for training purposes, reportedly worth $60M annually. The deal comes as the company seeks to increase its value before going public. This move is a departure from the previous practice of scraping data from the web without permission. Other companies, like Apple, have also sought deals with websites to train their AI models, with Apple reportedly negotiating multiyear deals worth at least $50 million with media companies. While Reddit’s terms and conditions may make this deal legal, it has sparked controversy among users, especially given Reddit’s recent decision to restrict access to its API.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| AI company strikes deal to use Reddit user content | AI training methods shifting from web scraping to deals with websites | More companies seeking deals with websites for AI training | Greater scrutiny of data scraping practices |
| Reddit restricts access to API to maximize revenue | Reddit limiting third-party app access for revenue | Increased monetization strategies by social media platforms | Maximizing revenue prior to IPO |
| Reddit users protest and subreddits go dark | User backlash against monetization strategies | Increased user activism and negotiation power | User opposition to monetization strategies |