ChatGPT gets custom instructions to remember your preferences

ChatGPT today needs no introduction. By far one of the most popular AI tools the world over, it has been a saviour to both, students and professionals alike. However, with the competition fast heating up on the artificial intelligence scene, OpenAI has been working on constantly upgrading its chatbot to stay ahead.

ALSO READ: xAI: Here’s what we know about Elon Musk’s latest venture

Therefore, a new feature, designed to ensure you have to use fewer inputs to get answers from the tool has been introduced to the chatbot, referred to as ‘Custom Instructions’.

What are custom instructions?

Custom instructions is a space wherein you can tell ChatGPT things about you, which the AI will keep in mind. This means, the answers to your queries will be in a manner that is tailored to you. All this without having to provide all the information and context over and over again.

ALSO READ: How to make ChatGPT and Google Bard forget you

Currently, each time one starts a new thread on ChatGPT, it does not remember past conversations, often providing generic answers. Now though, users can set custom instructions. For instance, setting tone for any email-related queries by default or limiting the answers to a specific number of words.

Who can use custom instructions?

As of now, this new feature is available as a beta version only to ChatGPT Plus plan users. It must also be noted that the feature is not yet available for users based out of the European Union and the United Kingdom.

Those eligible, can try out custom instructions on the desktop or laptop version by clicking on their name and going to settings from there. From there, one needs to select the “New Features” option and turn on the “Custom Instructions” setting. On iOS, users can go to “Settings,” and “New Features” from thereon, and then select “Turn on Custom Instructions.”

As per OpenAI, the information from one’s use of custom instructions, will help improve the AI model’s performance, teaching it to take a call regarding when to adapt its responses according to custom instructions, and when not to. However, bearing in mind its data privacy implications, users also have the choice to opt out of this, through the chatbot’s data control settings.

OpenAI says that it will use its moderation API, to scan instructions and check if they are unsafe in nature. ChatGPT has been given the ability to refuse to save the instructions and/or ignore them, if they may cause responses to violate the company’s policies. The same has been made as a safeguard to ensure users don’t type in instructions that may spread any misinformation or hate messages from ChatGPT.