close
close

What happens if your ChatGPT transcripts are leaked?

Data collection is once again at the forefront of new technology.

Illustration from The Atlantic. Source: Getty.

That is Atlantic Intelligence, a newsletter where our writers help you learn about artificial intelligence and a new machine age. Register here.

Shortly after Facebook became popular, the company launched an advertising network designed to allow companies to collect data about people and target them with marketing. So many problems in the social media age of the internet were due to this original sin. With this technology, Facebook, now Meta, would make its fortune and become dominant. And this is where our perception of online privacy changed forever, as people became accustomed to having different parts of their identities exploited and exploited by political campaigns, companies with something to sell, etc.

AI may change the way we experience the internet, but it is unlikely to turn back the clock on the so-called surveillance economy that defines it. In fact, my colleague Lila Shroff recently declared in an article for The AtlanticChatbots can only speed up data collection.

“AI companies quietly accumulate enormous amounts of chat logs, and their data policies generally allow them to do whatever they want. This can – what else? – mean advertising,” writes Lila. “So far, many AI start-ups, including OpenAI and Anthropic, have been cautious about advertising. But these companies are under great pressure to prove that the many billions in AI investments will pay off.”

Ad targeting may be unavoidable—since Lila wrote this article, Google has actually started rolling out such ads in some of its AI overviews—but there are other issues to contend with here. Users have long conversations with chatbots and often share sensitive information with them. AI companies have a responsibility to keep this data secret. But as Lila explains, there have already been glitches that leaked information. So think twice about what you type in the text box: you never know who will see it.

A silhouette that does that "silence" Gesture with a robot hand
Illustration from The Atlantic. Source: Getty.

Pssst, ChatGPT. That's a secret.

By Lila Shroff

Last spring, a man in Washington state feared his marriage was on the verge of collapse. “I'm depressed and going a little crazy, still love her and want to win her back,” he typed into ChatGPT. Using the chatbot, he planned to write a letter protesting her decision to file for divorce and post it on her bedroom door. “Highlight my deep guilt, shame and remorse for not taking care of her and not being a better husband, father and provider,” he wrote. In another message, he asked ChatGPT to write his wife a poem “epic enough to make her change her mind, but not cheesy or over the top.”

The man's chat history was included in the WildChat dataset, a collection of 1 million ChatGPT conversations collected consensually by researchers to document how people interact with the popular chatbot. Some conversations involve requests for marketing texts and homework help. Others might feel like you're peering into the living rooms of unsuspecting strangers.

Read the full article.

What to read next

P.S

Meta and other companies are still trying to make smart glasses a reality — and generative AI could be the secret ingredient that makes the technology click, my colleague Caroline Mimbs Nyce wrote in a recent article. What do you think: would you wear them?

–Damon

You may also like...