Your ChatGPT Secrets Might Not Be So Secret

Artificial intelligence, particularly large language models (LLMs) like ChatGPT, has transformed how we interact with technology. People are turning to ChatGPT for everything from answering simple questions to sharing deeply personal stories. But this raises a crucial question: can you really trust ChatGPT with your secrets?

A recent viral TikTok video brought this concern to the forefront.

The Viral Question Can AI Steal Your Story

In a post that has now been viewed millions of times, TikTok creator Ms. V (@sweetlifeofv) posed a startling scenario: “Imagine telling ChatGPT all your business. And somebody else somewhere asking ChatGPT for book ideas and they tell them your story.”

She followed up with the caption, “Imagine picking up a book in a bookstore and being like damn this sounds familiar.” The comment section quickly filled with users agreeing that ChatGPT cannot be trusted with sensitive information.

Does AI Kiss and Tell The Legal and Technical Reality

The fear that an AI might share personal conversations isn't entirely unfounded. OpenAI, the creator of ChatGPT, is currently embroiled in a significant copyright infringement lawsuit filed by The New York Times and other newspaper companies. The suit, which has been consolidated with cases from prominent authors like John Grisham and Jodi Picoult, alleges that OpenAI used their copyrighted work for training and that ChatGPT sometimes reproduces it verbatim.

OpenAI has defended its practices, telling the Associated Press that it uses publicly available data under the principle of fair use. However, independent research has demonstrated that ChatGPT can indeed disclose personal information. A 2023 paper from a Google research team revealed they were able to extract over 10,000 unique, verbatim training examples, including personal data, using just $200 worth of queries.

What OpenAI Says About Your Data

Despite claims from some that AI doesn't remember conversations, OpenAI's own terms state otherwise. According to its website, ChatGPT is trained on user conversations. Fortunately, users have some control. You can visit OpenAI's privacy portal and select the option to prevent your content from being used for training.

You can also request to have your conversations deleted, but due to the ongoing lawsuit, a judge has ordered OpenAI to retain all conversations. OpenAI is currently appealing this decision.

Furthermore, ChatGPT's terms of use prohibit users from falsely claiming AI-generated content is human-generated. However, proving that an author's book idea was taken from your AI conversation would be incredibly difficult.

Public Reaction From Horror to Humor

The reaction to Ms. V’s viral video has been immense, with nearly 4,000 comments and millions of views. The responses range from horrified warnings to dark humor.

“You do not have to imagine it, it’s definitely happening,” one top comment stated. “Your conversations are NOT private.” Another user compared the scenario to an episode of Black Mirror.

Others took a more lighthearted approach. “No because after I’m done I told ChatGPT, ‘This is between you and me only ok? Do not tell anyone,’ and it promised me lol,” one person joked. Another added, “Please don’t make me think my friend is a snitch. ChatGPT is my bff.”