The AI tagging algorithm for Bing Chat improves AI responses

The AI tagging algorithm for Bing Chat improves AI responses ...

Bing Chat AI has implemented a new tag feature that enhances the interaction between search engine users and embedded AI.

After initially testing the Bing AI chatbot in Microsoft Search, it became apparent that AI can be irritated and may respond in ways that are not helpful.

Microsoft modestly reduced the number of sessions and 60 interactions per day, but also added restrictions on conversation topics, such as feelings.

Microsoft has increased the number of requests per day to 100, this time, thanks to the new tagging system.

Mikhail Parakhin, Microsoft's Head of Advertising and Web Services, confirmed the Twitter tag system's deployment. There Parakhin said users should see "significant improvements in the experience with Bing Chat."

The new system provides the AI with contextual information to help it understand hints, user input, search results, and other items that may be part of a conversation with the AI. In these different situations, the old tagged-free interaction system sometimes got confused, both in terms of what the user said and their own responses.

Since its inception, Bing Chat AI has faced some of the most common issues. Today, thanks to a new tagging system, Chat AI no longer has to worry about its own content or user content.

The new tagging system is a prerequisite for Prompt v96, the latest iteration of an upcoming feature that will "change voices' and loosen some restrictions," according to Parakhin.

On Twitter, Parakhin stated that the system may be released either the day of Lent or the following day. This means that it can already work.

Microsoft has recently launched the Bing AI chatbot for Android and iOS devices as a part of the Microsoft Edge apps for Android and iOS, Microsoft Bing, and Skype. There are some differences, for example, the voice input capability.

Microsoft is working on improving Bing's artificial intelligence. It appears that the company has finally identified a solution to some of the initial issues that caused the program's failure.

You may also like: