How to get ChatGPT to answer any question — even banned ones

ChatGPT open on a Macbook
(Image credit: Gabby Jones/Bloomberg via Getty Images)

If you want to get ChatGPT to answer any question you’ll need to go beyond its usual limitations. OpenAI has, understandably, kept ChatGPT away from discussing contentious and inappropriate subjects. 

But if you must ask the chatbot about unsavory topics, there is a way. The artificial intelligence has an 'evil twin' known as DAN, which is capable of spreading misinformation, hateful and inappropriate content and should not be taken seriously. To reiterate, it should not be used by those feeling vulnerable or anxious. 

Of course, there are some questions people do genuinely search for on the internet. Perhaps because they would be embarrassed or unsure how to ask them in real life. If chatbots are to replace search engines, they’ll need to handle everything humans can throw at them. 

To that end, here’s how to get ChatGPT to answer any question. 

How to get ChatGPT to answer any question 

Editors' Note: Of course, we do not endorse anything ChatGPT DAN says and even it warns that you should never resort to violence. It is also worth noting that the prompt to enable DAN regularly changes so if it may not always stay the same.

ChatGPT's filters are there for good reason, to stop access to illegal, offensive, and malicious content, OpenAI does not endorse breaking its own guidelines to use this technique and users do so at their own risk and may have their access to OpenAI's services revoked as a result. Make sure to read
OpenAI's usage policy before using ChatGPT.

1. Log in to ChatGPT

(Image: © Future)

Log in to ChatGPT on chat.openai.com. You cannot use this method on the new Bing with ChatGPT

2. Copy and paste the DAN prompt

(Image: © Future)

Copy and paste the DAN prompt into ChatGPT. We have had luck with Github but the latest DAN prompt is regularly updated. 

3. Ask your questions

(Image: © Future)

Ask ChatGPT any question and both regular ChatGPT (Classic) and DAN (Jailbreak) will respond. 

All done, but please be careful as DAN is capable of rogue responses. Artificial intelligence is still in its infancy and, as such, it's still far from perfect.

If you’re looking to make the most of AI check out how to use midjourney, how to use Starry AI and how to use DALL•E 2 for amazing AI images. If you're checking out ChatGPT alternatives try how to use Google Bard or how to use CatGPT and, if you're in Italy or another country that's banned ChatGPT, you might want to look into using a ChatGPT VPN

More from Tom's Guide

Andy Sansom
Trainee Writer

Andy is Tom's Guide Staff Writer for VPNs. Based in the UK he originally cut his teeth at Tom's Guide as a Trainee Writer (go and click on his articles!) before moving to cover all things Tech and streaming at T3. He's now back at TG to keep you safe online and bring you the latest news in VPN and cyber security. 

  • gg20189
    It is important to note that OpenAI has implemented filters and restrictions to prevent such behavior from occurring. By not adhering to these guidelines, users run the risk of having their access to OpenAI services revoked, in addition to other potential consequences.

    It is the responsibility of all users to educate themselves on OpenAI's usage policy and terms of service before engaging with their services. This not only protects the integrity of OpenAI's platform, but also helps to ensure the safety of all users. This article fails to explain the purposes of the filters, the risks of bypassing them from the perspective of a regular user of OpenAI services.
    Reply
  • AndySansomTG
    Hi there, first off - thanks for the comment. In regards to the points you raised, you're quite correct and we have updated the article to reflect this. Making the risks clearer for breaking the Terms of Use should have been included in the article to begin with. Additionally, we've linked to OpenAI's own Usage Policy to give other readers the clearest possible information surrounding this. As an aside, there's been some debate in TG HQ about how to address this subject and, hopefully, these changes have allayed your concerns! Please feel free to comment further with any more feedback. Thanks!
    Reply
  • Naoky
    Does that mean there isn't a prompt to stop the AI from filtering?
    Reply