US DOJ Backs OpenAI: Training LLMs on Copyrighted Content Can Be 'Fair Use'

What happened

In the New York Times v. OpenAI copyright case, the US Department of Justice filed a brief supporting OpenAI, arguing that training large AI models on copyrighted content can constitute fair use.

Significance

• The brief does not settle the lawsuit, but sends a powerful political signal: Washington views broad AI training rights as strategically important
• If courts adopt this view, it would reshape how AI companies obtain training data across the industry

Potential impact

• Positive: lower legal risk for model makers over training data, more stable cost expectations
• Tension: content owners (news, publishing, stock photography) lose licensing leverage, making future data-licensing negotiations more complex

Source: The Automated Daily / Bloomberg