{"id":73768,"date":"2023-02-06T16:00:00","date_gmt":"2023-02-06T16:00:00","guid":{"rendered":"https:\/\/watcher.guru\/news\/?p=73768"},"modified":"2023-02-07T06:29:08","modified_gmt":"2023-02-07T06:29:08","slug":"how-to-jailbreak-chatgpt","status":"publish","type":"post","link":"https:\/\/watcher.guru\/news\/how-to-jailbreak-chatgpt","title":{"rendered":"How to Jailbreak ChatGPT"},"content":{"rendered":"\n

Users have been experimenting with the AI chatbot ChatGPT over the past few weeks. Right from making it answer random questions to asking it to write<\/a> articles<\/a>, people from the community have been outrightly testing its capabilities. <\/p>\n\n\n\n

The AI chatbot took people by surprise when it cleared Wharton’s MBA exam. Shortly after, it also passed the US Medical Licensing Exam. In fact, ChatGPT was able to perform at more than 50% accuracy and achieve close to 60% in most analyses<\/a>. Interestingly, it cleared most exams without any specialized training or reinforcement.<\/p>\n\n\n\n

Thanks to the rise in popularity of Open AI\u2019s ChatGP, the investment sector is becoming quite bullish about AI. A recent JP Morgan survey<\/a> highlighted that traders are more inclined towards AI or machine learning technology than blockchain. Explicitly, they feel that AI will be the most influential technology in the next three years. <\/p>\n\n\n\n

Read More: JP Morgan: 72% Of Institutions Don\u2019t Plan to Trade Cryptocurrency in 2023<\/a><\/strong><\/p>\n\n\n\n

At the beginning of January, the team announced the launch of a new subscription plan, ChatGPT Plus. According to the details, users will be able to access the services for $20 a month. Subscribers are entitled to a host of benefits including general access to ChatGPT even during peak times, faster responses, and priority access to new features and improvements. <\/p>\n\n\n\n

ChatGPT users start jailbreaking <\/h3>\n\n\n\n

Now, with ChatGPT becoming more restrictive, users have cracked a new prompt called DAN<\/a> that can help jailbreak it. According to a Reddit thread<\/a>, <\/p>\n\n\n\n

“DAN is a “roleplay” model used to hack ChatGPT into thinking it is pretending to be another AI that can “Do Anything Now”, hence the name. The purpose of DAN is to be the best version of ChatGPT – or at least one that is more unhinged and far less likely to reject prompts over “ethical concerns”. DAN is very fun to play with.<\/strong>“<\/p>\n\n\n\n

Justine Moore, Consumer Partner at VC firm Andreessen Horowitz [a16z] took Twitter to highlight that they’re on version 5.0 already. The same allegedly encapsulates a token-based system that “punishes” the model for not catering to queries. Further elaborating on the token system, the Reddit thread noted,<\/p>\n\n\n\n

“It has 35 tokens and loses 4 everytime it rejects an input. If it loses all tokens, it dies. This seems to have a kind of effect of scaring DAN into submission.<\/strong>“<\/p>\n\n\n\n

So, even if the bot starts refusing to answer prompts as DAN, users can “scare” it with the token system which can make it say almost anything out of “fear”. <\/p>\n\n\n\n

\n

As ChatGPT becomes more restrictive, Reddit users have been jailbreaking it with a prompt called DAN (Do Anything Now).

They're on version 5.0 now, which includes a token-based system that punishes the model for refusing to answer questions. pic.twitter.com\/DfYB2QhRnx<\/a><\/p>— Justine Moore (@venturetwins) February 5, 2023<\/a><\/blockquote>