OpenAI asked a federal judge in New York on Wednesday to reverse an order that required it to turn over 20 million anonymized ChatGPT chat logs amid a copyright infringement lawsuit by the New York Times (NYT.N)
The artificial intelligence company argued that turning over the logs would disclose confidential user information and that "99.99%" of the transcripts have nothing to do with the copyright infringement allegations in the case.
"To be clear: anyone in the world who has used ChatGPT in the past three years must now face the possibility that their personal conversations will be handed over to The Times to sift through at will in a speculative fishing expedition," the company said in a court filing.
LAWSUIT CLAIMS OPENAI MISUSED TIMES STORIES
The news outlets argued that the logs were necessary to determine whether ChatGPT reproduced their copyrighted content and to rebut OpenAI's assertion that they "hacked" the chatbot's responses to manufacture evidence. The lawsuit claims OpenAI misused its articles to train ChatGPT to respond to user prompts.
Magistrate Judge Ona Wang said in her order to produce the chats that users' privacy would be protected by the company's "exhaustive de-identification" and other safeguards. OpenAI has a Friday deadline to produce the transcripts.
OpenAI Chief Information Security Officer Dane Stuckey said in a blog post on Wednesday that sharing the logs would violate privacy and security protections and "force us to turn over tens of millions of highly personal conversations from people who have no connection to the Times’ baseless lawsuit."
A New York Times spokesperson said OpenAI's blog post "purposely misleads its users and omits the facts."
"No ChatGPT user’s privacy is at risk," the spokesperson said. "The court ordered OpenAI to provide a sample of chats, anonymized by OpenAI itself, under a legal protective order."
The OpenAI case is one of many pending lawsuits against tech companies over the alleged misuse of copyrighted work to train AI systems.
ChatGPT has received yet another big upgrade, this time in the form of GPT-5.1. OpenAI claims the update will make it "easier to customize" the chatbot, while also providing a "smarter, more conversational" experience across the board.
The company revealed the release of GPT-5.1 in a blog post today, noting that it will have two new models called GPT-5.1 Instant and GPT-5.1 Thinking. Instant is meant to provide a "warmer, more intelligent" model that is better at following the user's instructions, while Thinking will focus on delivering advanced reasoning. OpenAI says that ChatGPT will not only be equipped to handle simple tasks more efficiently, but that the AI will also have a better grasp of complex tasks.
The company provided some examples of how outputs from GPT-5 and GPT-5.1 Instant will differ, and there is definitely a bit more of what the company calls "playfulness" visible in the messaging. Additionally, the company says that GPT-5.1 Instant will use adaptive reasoning to determine if it should think before responding to more difficult questions, as this will help it provide more accurate answers without bogging down the system too much.
Instant isn't the only model that will provide a warmer response, either, as OpenAI says Thinking will offer a warmer and more empathetic response system, even showcasing the difference between it and GPT-5. But the real kicker here is that the company claims this "release is a step forward" for the models, not only in capability but also in usability. And a big part of that is because of the new personalization features.
Beyond just adding support for GPT-5.1, you can also now further customize how ChatGPT responds to you based on your personality. The examples offered in OpenAI's blog post show options like Candid, Quirky, Efficient, Professional, Friendly, and more. Once set, the personality will be utilized across all models of the AI. The previous Nerdy and Cynical options will remain available, too.
The company is also toying with more granular approaches to customization, which could see even more fine-tuned settings offered in future releases, akin to the custom instructions we can already give the AI. There wasn't any mention of the recently leaked ChatGPT group chat feature, so we'll have to wait to learn more about that one. Auto will continue to let you rely on the AI to determine which model is best suited to a task. Additionally, if the warmer return from GPT-5.1 is too much, you can always use GPT-5 through the legacy models option in ChatGPT, the same way that you used GPT-4 when GPT-5 came out.
OpenAI
I've been a longtime ChatGPT Plus user and talk to the AI chatbot regularly. Whether it's for work or fun, ChatGPT can be incredibly helpful as long as you're aware it can hallucinate. It's never a bad idea to verify its claims, especially if you're concerned it might be wrong.
I've also been keeping up with all the new features that OpenAI has launched in recent years and testing them accordingly. That includes all the UI changes that came with the various novelties that OpenAI added to ChatGPT.
On that note, I'll remind you that ChatGPT Search and the addition of Advanced Voice Mode and Canvas made OpenAI deliver several big changes to the ChatGPT UI, which makes the tool even easier to use.
However, to get the best results, you need to be mindful of one thing: the prompt. That is your first set of instructions in a conversation. The better the prompt, the faster you get the desired results. Otherwise, you must keep tweaking it for the chatbot to give you what you need.
This is where I tell you ChatGPT has a built-in "hack" to improve your prompts, one that I've missed for many months. It's called "Edit message," and it could have saved me hours of prompt editing in all my chats with ChatGPT.
Whenever I start a new conversation with ChatGPT, I try to provide as many details and instructions as possible in the prompt to prevent wasting time with follow-up requests.
Sometimes, I get it right on the first try, but many times, I have to tweak my initial request. To fix the prompt, I write a reply in the chat, asking the AI to tweak its response to take into account additional requests.
I was testing ChatGPT's new Tasks feature on Wednesday when I realized I'd been making a huge mistake for months. I ignored the "Edit message" button that appears when you hover over one of your replies. The editing option is available in the mobile app if you touch and hold any of your prompts.
We've been editing replies in instant messenger apps like iMessage and WhatsApp for years. I'm more than familiar with the feature. It's a great way to fix an error in a chat reply in a text with other people. Of course, you should expect AI chatbots like ChatGPT to offer the same functionality.
Why is the "Edit message" so brilliant? Suppose you start a new chat with ChatGPT, and your initial prompt doesn't yield the desired results. You could tap that button and edit your initial prompt rather than writing a reply to the AI to tweak it. ChatGPT will adapt to the changes automatically and provide a different answer.
The best part is that ChatGPT will show you the previous response so you can choose how to continue your conversations.
There's another benefit to using the button. The conversations won't always be as long, so you'll have less scrolling to do when revisiting a chat. Also, you can use the button as many times as you need to get the proper start.
The screenshots above show an example of editing a first prompt in a ChatGPT conversation. I wrote a short prompt, waited for ChatGPT to respond, and then edited the prompt to something entirely different. As you can see in the screenshot sequence, ChatGPT adapted accordingly.
The "Edit message" button is available for all the user's replies in a chat, not just the first one.
While I missed using this ChatGPT prompt "hack" right before my eyes for months, I'll surely take advantage of it now that I've realized how bad my ways were.