Microsoft Rolling Out AI-Powered Bing Generative Search, Could Rival Google’s AI Overviews

Microsoft added a new artificial intelligence (AI) feature to its Bing search engine on Wednesday. Dubbed Bing generative search, the feature shows a snapshot with the information for the searched query. The feature is similar to Google’s AI Overviews which was publicly released in the US in May. Bing’s AI-powered search feature also shows the sources from where the content was taken. As per the company, the feature is currently being shipped for a small percentage of queries, likely to avoid incidents of AI hallucinations that AI Overviews suffered from.

Bing Generative Search Features

In a blog post, Microsoft Bing announced the feature and said, “Today, we’re excited to share an early view of our new generative search experience which is currently shipping to a small percentage of user queries.” The feature is an extension of the company’s AI-powered chat answers on Bing which was released in February 2023.

Bing generative search
Photo Credit: Microsoft Bing

The new feature is built with the company’s large and small language models (SLMs and LLMs), however, the company did not reveal which AI models were used for it. Explaining the functioning, Microsoft highlighted that the AI can contextually understand the search query, and then processes millions of sources of information to match the right content. After that, it uses the data to generate relevant search results.

Visually, Bing generative search looks very similar to Google’s AI Overviews. It appears inside a grey box at the top of Bing search result page. The snapshot of information comes with a headline, a document index for easy navigation to the right section, and source citations. The AI also adds images, videos, and tables whenever necessary.

Whenever the Bing generative search shows up, the traditional search results will move to the right side of the page. The left side first shows the snapshot, followed by the source links, and then related search results.

The company says it is optimising accuracy and has refined its approach to focus on that. However, just like the details around the AI model, the architecture used for this was not disclosed. Microsoft also highlighted that it is closely monitoring how the feature impacts traffic to publishers.