ChatGPT users will be getting a customization option – Technology Org

OpenAI, the company leading the development of the globally-known ChatGPT artificial intelligence chatbot, announced an upcoming new upgrade to its product that allows users to customize it.

ChatGPT from OpenAI on a smartphone - illustrative photo

ChatGPT from OpenAI on a smartphone – illustrative photo. Image credit: Focal Photo via Flickr, CC BY-NC 2.0

According to the developer, the reason for such a feature upgrade is its aim to fight the bias in artificial intelligence.

The company has noted it is working on ways to diversify views presented in the content generated by ChatGPT, which should also mitigate different biases – for example, political ones.

“This will mean allowing system outputs that other people (ourselves included) may disagree strongly with,” the company said in its blog post.

The existing functionality of ChatGPT already allows for the automatic creation of human-like responses to very varied user requests. The platform was officially released last November, but public interest in this product was immense – up to the point where schools and universities started banning the use of ChatGPT in students’ academic activities.

This month, some media sources have also pointed out that certain answers generated by chatbots based on technology from OpenAI “may be dangerous”.

OpenAI is currently cooperating with Microsoft on the integration of their technology in the Edge internet browser. Microsoft announced on Wednesday, that user feedback is a crucial component needed to improve generative algorithms, and a lot of work is being done with the aim to ensure that there was no possibility to “provoke” AI to generate non-intended responses.

According to OpenAI, its ChatGPT is trained on large datasets of human-created content, but only after human moderators review these datasets according to given guidelines on how to respond in different situations – especially when these scenarios are complicated.

In its existing version, ChatGPT has certain hard-coded safeguards against adult content, and violent or hate speech. In these situations, the AI ​​platform is directed to provide answers similar to “I can’t answer that.”

If the topic is deemed controversial, ChatGPT tries to avoid biases by describing the viewpoints of people and movements, but does not assume a very specific position.

Leave a Reply

Your email address will not be published. Required fields are marked *