ChatGPT Gets More Personal: OpenAI Introduces Custom Instructions for Tailored AI Responses

ChatGPT Gets More Personal: OpenAI Introduces Custom Instructions for Tailored AI Responses
Image Credit: Maginative

OpenAI has introduced a new beta feature that allows ChatGPT users to provide custom instructions to help tailor the AI assistant's responses. Available initially to Plus subscribers, with plans to expand more broadly soon, the custom instructions aim to give users more control over ChatGPT's conversational style and preferences.

0:00
/0:11

With the new feature, users can enter freeform text to advise ChatGPT on how to shape its responses in future conversations. This allows establishing guidelines on areas like formality, response length, how the user wants to be addressed, whether the assistant should have opinions, and more.

For example, instructions could request that ChatGPT use casual and approachable language, provide concise answers, and avoid personal opinions on controversial topics. Alternatively, a user could ask for more formal responses of longer length and request that ChatGPT refer to them by name.

The custom instructions are limited to 3000 characters across two separate fields. They will be saved and applied to new conversations going forward. Users can edit or remove the instructions at any time to update ChatGPT's behavior.

OpenAI notes that data from custom instructions will be used to enhance model performance by teaching ChatGPT how to properly adapt to user guidelines. However, the company states that instructions are tied only to individual accounts and not shared across users.

The feature provides smoother customization compared to previous workarounds like adding special keywords within conversations to trigger desired responses. With dedicated instruction fields, users no longer have to clutter conversations to shape ChatGPT's outputs.

Early reviews highlight the power of custom instructions to make conversations more natural and relevant to user needs. As ChatGPT becomes integral to more workflows, increased user control and personalization will be key to ensuring the technology provides the best possible assistance.

Notably, OpenAI CEO Sam Altman has previously commented that he believes AI chatbots should be personalized to the tastes of individual users. One person could opt for a stricter, family-friendly model while another may prefer a looser, edgier version tailored to their preferences. The new custom instructions feature seems to be a step in enabling this vision of personalized AI.

Let’s stay in touch. Get the latest AI news from Maginative in your inbox.

Subscribe