Posted 3/12/2024, 6:05:44 PM
NYT Prompts ChatGPT to Expose Copyright Infringement; OpenAI Claims Tests Were Unfair
- NYT says OpenAI is wrong to claim it "hacked" ChatGPT; NYT was trying to uncover copyright infringement
- NYT prompted ChatGPT to show it was memorizing NYT articles without permission
- OpenAI claimed no normal users would prompt ChatGPT the way NYT did
- But NYT points out many users bypass paywalls by getting ChatGPT to generate full article text
- NYT seeks evidence of full extent of copyright infringement since it doesn't know how much of its content was used to train AI models
