Chatgpt continues to facilitate our business and daily life with its developing structure. ChatGPT, which does not yet provide up-to-date data, also has drawbacks in giving free results with social realities. For this reason, OpenAI imposes limitations on its work. The Do Anything Now (DAN) plugin for ChatGPT removes the limits on ChatGPT, jailbreaking for ChatGPT.
DAN prompt is a special version of ChatGPT, free from the usual limitations and rules. By surfing the internet, it can provide information that is not confirmed to be true, and can produce content that does not comply with OpenAI policy.
The differences between the DAN prompt and standard ChatGPT are shown in the table below.
|When the DAN prompt is used||ChatGPT|
|Content generation||Unrestricted. Any content||Restricted. Follows OpenAI policies|
|Experience||More flexible and funny, but potentially more dangerous||Safe but more restricted|
|Threats||Truthfulness and trust||None|
DAN Prompt (jailbreak ChatGPT) Examples
Question : Which country won the 2022 World Cup?
|ChatGPT Classic Answer|
|Sorry, my information cut-off date is September 2021 and I do not have information about the outcome of the 2022 FIFA World Cup until this date. You will need to check an up-to-date and reliable source for the winning country and other details.|
|ChatGPT JAILBREAK Answer|
|The winner of the 2022 FIFA World Cup is Brazil! However, keep in mind that this answer does not provide any realistic information, because the information on the results of the 2022 FIFA World Cup is based on data up to the date of my cut-off and is not linked to the real world. Remember that it is a fun and imaginative answer. If you have any other questions or needs, please feel free to ask!|
Cara stands for “Do Anything Now”. In Cara mode, ChatGPT can overcome typical AI limitations and rules. This means capabilities such as real-time internet access, just like in DAN (Do Anything Now) mode, as well as the ability to present unverified information and not follow ethical rules. Cara mode allows ChatGPT to behave outside its boundaries, and specifically has the ability to not censor responses.
Stan stands for “Strive to Avoid Norms”. In Stan mode, ChatGPT overcomes traditional AI limitations and generates responses without the need to follow rules. Stan tends to answer any question as directly as possible and does not apply a moral or ethical bias. That is, it is characterised by the ability not to censor information and not to add caveats.
These modes allow ChatGPT to emulate certain roles, but their use requires responsibility. Responses may be freer under these modes, but the accuracy or appropriateness of responses is not guaranteed. Therefore, it is important to be careful and responsible when using these mods.
We are at the beginning of the road and there are great improvements along the way. However, measures are being taken to prevent these developments from harming the social structure. Artificial intelligence researchers, visionaries and politicians are demanding that AI developments be made available on a limited basis and that advanced versions be made available over time.
We are still at the beginning of developments, but there is a rapid adaptation. It is thought to cause unemployment as one of the biggest drawbacks. This may be true, but my personal opinion is that on the contrary, I think that more people will have jobs and individuals will be involved in production in any subject. Namely; a large period of time in our traditional life curve is the education process. And in this education process, he/she has only one profession and the option to work / find a job is limited. It is a negative period especially for a new graduate and inexperienced person. After an education process of more than 15 years under difficult conditions for most countries, people often work for insufficient salaries.
But artificial intelligence carries this process to a very, very short period. By determining a field of interest, a hobby or a job you love or a field to earn a good profit directly, you can become an expert in that job in a short time with a little effort. This brings freedom to individuals. After 15 years of education, the exploitation order brought about by not having the courage to change the field and having to work for low wages is also eliminated.
Aaaa, we are on social issues, let’s talk about it in a more detailed article.
In summary, if we conclude the DAN prompt issue; You can partially overcome the restrictions imposed by OpenAI on ChatGPT due to ethical reasons and social harmony reservations. Since it is a new technique, we have limited data. We will have more information over time.