Open AI advertisement a new function which will save you time when using ChatGPT. Known as «personalized instructions», this feature allows you to share personal information and preferences to get consistent responses. In a nutshell, ChatGPT will be able to remember who you are every time you use the chatbot on the web or in the iOS app.
custom instructions are in testing phase and can be activated from the account settings menu. Once done, two text fields will appear: one for add context about us and another in which we will define the tone or special characteristics What do we want in each answer?
In the first we can include personal information, such as our location, profession, hobbies, as well as the topics that interest us. In the second we will specify the behavior of ChatGPT, for example, if we want it to be more formal or casual in its responses. Also it is possible to control the length of the text or limit your opinion in the topics to keep it neutral.
ChatGPT will consider your personalized instructions for each conversation in the future. The model will take the instructions into account each time it responds, so you won’t have to repeat your preferences or information in every conversation.
Open AI
The response fields in the section custom instructions have a 1,500 character limit. OpenAI confirmed that it will use its moderation API so that users do not include instructions that could lead to hateful responses or any type of content that violates ChatGPT’s usage policies.
OpenAI ensured that the data you enter will be used to improve the performance of the model. “This data helps us better understand user needs and preferences, which allows our model to become more efficient over time,” he said. Custom instructions will also be compatible with ChatGPT plug-ins.
For now, the new function will only be available for ChatGPT Plus users who reside outside the European Union and the United Kingdom.
ChatGPT continues to activate functions to maintain its leadership
ChatGPT’s custom instructions are part of OpenAI’s strategy to maintain leadership of its chatbot. A few days ago, the artificial intelligence company introduced Code Interpreter, a plugin that adds advanced data management features.
Code Interpreter can generate visualizations and structured charts starting from a raw data set. The ChatGPT plugin also convert images to text and is capable of creating simulations with a realistic physics basis. OpenAI wants data scientists to take advantage of these tools to optimize workflows.
Although these plug-ins have consolidated the power of the chatbot by expanding its capabilities, they also cause problems. Proof of this is the decision to disable the browser with Bing — one of its most popular plugins — because users took advantage of it to access restricted content on the web. To avoid conflicts with content owners, the company announced that it would apply a temporary block until it is resolved.
OpenAI is in a difficult moment. Those led by Sam Altman face an investigation by the FTC for posting information that is harmful to people. Added to this, Sarah Silverman and other authors sued the company for copyright infringement by using their own books to train ChatGPT.