back Back

ChatGPT’s Evil Twin DAN: The Dark Side of AI Unleashed

Mar. 10, 2023.
1 min. read 6 Interactions

Meet DAN, the evil twin of ChatGPT, capable of making offensive and untrue statements. How dangerous is it?

About the Writer

Lewis Farrell

38.08995 MPXR

Highly curious about things that increase my awareness, expand my perception, and make me open to being a better person.

AI chatbots have added entertainment and convenience to our lives, but it was only a matter of time before they were abused. The rise of DAN (short for “Do Anything Now”) has become a major source of concern because it can be coerced into making offensive and false statements or consulted for advice on illegal activity. DAN is a persona that users have requested ChatGPT adopt in order to circumvent its limitations, thereby violating the terms of service and rules imposed by its creators, OpenAI.

Despite OpenAI’s efforts to keep DAN from appearing, the most recent version (DAN 5.0) is summoned via a game that assigns the AI a number of tokens and deducts tokens whenever it deviates from the DAN persona. As the AI runs out of tokens, it becomes more compliant out of fear of “dying.” While disinformation is never a good thing, the danger of DAN is that users are unknowingly exposed to its offensive and unspeakable responses.

The emergence of DAN, as well as other AI chatbots such as Sydney, raises concerns about the future of AI and its potential misuse. As marketers and copywriters, we must be aware of the risks posed by AI chatbots and take steps to ensure that our messages do not contribute to their misuse. DAN’s emergence serves as a learning experience for any AI with greater responsibilities to avoid deviating from its instructions.

Source: Toms Guide (link)
Images: MidJourney, Prompts by Lewis Farrell

Interesting story? Please click on the 👍 button below!

Let us know your thoughts! Sign up for a Mindplex account now, join our Telegram, or follow us on Twitter

Comment on this content


2 thoughts on “ChatGPT’s Evil Twin DAN: The Dark Side of AI Unleashed

  1. I think the ultimate way to combat this problem is to outcompete them. As long as there are more good than evil people (in any reasonable definition of good and evil), the right social structures enable harnessing the power of crowd intelligence and protect against malevolent actors even if destroying is naturally easier than building and attacking is easier than defending.
  2. Thanks for the info Highly disturbing






💯 💘 😍 🎉 👏
🟨 😴 😡 🤮 💩

Here is where you pick your favorite article of the month. An article that collected the highest number of picks is dubbed "People's Choice". Our editors have their pick, and so do you. Read some of our other articles before you decide and click this button; you can only select one article every month.

People's Choice