Why Did the New York Times Sue OpenAI?
Lessons learned from The Times lawsuit over AI
You may have heard that the New York Times is suing AI juggernaut OpenAI about its alleged misuse of content. Why?
In their lawsuit, The Times shared Exhibit A below. The left column shows ChatGPT’s output in response to a prompt about Apple Computer entered by New York Times lawyers. On the right is the original article. 97% of the answer — 215 of 222 words — are the same.
Drop the mic! This is obvious and irrefutable evidence that OpenAI is ripping off copyrighted New York Times material!
Not so fast.
Is the New York Times telling the whole story?
OpenAI’s response to The Times is fascinating. It explains how they’re collaborating with publishers, how training with public data is fair use, how they allow any publisher to opt out of its content being used to train AI, how “regurgitation” like this is a rare bug they’re working to eliminate, and how stolen content usually comes from third-party sites, not www.nytimes.com.
But the fourth section, “The New York Times is Not Telling the Full Story,” is the most interesting.
Regurgitation or manipulation?
According to OpenAI, The Times appears to have purposefully manipulated the questions it posed ChatGPT to regurgitate copyrighted content, then cherry-picked examples to influence the court of public opinion.
OpenAI’s claim seems viable — it’s easy to instruct AI to violate copyrights. AI engines are like a golden retriever: they’ll fetch whatever you tell them to.
For example, ask an AI image-generator like Midjourney to “imagine” a Joaquin Phoenix Joker movie, 2019, a screenshot from a movie, or a movie scene, and you’ll get what you ask for:
Midjourney can be coerced to infringe copyrights — who is responsible?
What’s going on here? Are the companies that make AI responsible for how you use it?
Who’s responsible for AI’s output?
The Times claims OpenAI should be held responsible for billions in damages and names Microsoft as a defendant (Microsoft has invested an estimated $10 billion in OpenAI).
The lawsuit claims “millions” of articles published by The Times were used without permission so readers could get their content from OpenAI without paying. The claimed damage comes from lost subscription and advertising revenue. OpenAI claims that this is because other parties stole it, and that regurgitation is rare and a bug they’ve engaged publishers to fix.
I think The Times lawsuit reveals something else.
Use AI as a copilot, but you’re flying the plane
The Times lawsuit shows who’s really responsible for AI: you. NEVER use the output of an AI engine straight out of the box without heeding established rules of creative use. Just like in the real world.
AI is rocket fuel for creativity. Tell it to fetch what you wish. Be inspired by Joaquin Phoenix and The New York Times, but don’t rip them off.
Use AI as a copilot, but you’re flying the plane.
This week, I’m publishing answers to questions from The Generative AI Growth Mindset workshop. It’s inspired firefighters, teachers, and students to jump in and use AI. It’s just $49, has a 9+ rating out of 10, and 100% of net proceeds support Hospice charity. So far, we’ve generated $2,800 while spreading the good word about how to use AI. Register for the next cohort here.
