- Microsoft today announced an important update to its AI chatbot: visual search.
- Users can now take or upload an image to Bing Chat and request more information about it via the desktop or the Bing app, where the AI can now “understand the context of the image, interpret it, and answer questions about it,” according to a statement.
- The update comes as the AI arms race rages on among chatbot leaders such as Microsoft, Google, OpenAI, and Anthropic.
Satya Nadella, CEO of Microsoft Corp. , during the company’s Ignite Spotlight event in Seoul on November 15, 2022.
Seung Joon Cho | bloomberg | Getty Images
Microsoft on Tuesday announced a significant update to its artificial intelligence chatbot: visual search. Users can now take or upload a photo to Bing Chat and request more information on it through desktop apps or Bing.
“Bing can understand and interpret the context of the image and answer questions about it,” Microsoft wrote in a statement. “Whether you’re traveling to a new city on vacation and asking about the architecture of a particular building or at home trying to come up with lunch ideas based on the contents of your fridge, upload the image to Bing Chat and use it to harness the knowledge of the web to get you answers.”
The update comes as the AI arms race rages on among chatbot leaders such as Microsoft, Google, OpenAI, and Anthropic. In an effort to develop the most advanced AI, the tech giants are rapidly launching new features, aiming to keep up with not only their text chat bot competitors, but also the AI heavyweights.
Although searching for images—and responses that include images—are now part of the user experience for chatbots, none of the leading text-based chatbots seem to be able to generate their own images yet, unlike tools like Midjourney, Stable Diffusion, and DALL- E 2. However, Google says the feature is on its way for Bard chatbot.
Microsoft’s decision to allow images for Bing Chat follows Google’s recent debut of an image search feature for Bard, its chatbot. With Google Lens, users can request information from the Bard about a photo they’ve uploaded, ask them to create a comment or even just add some zest to chatbot responses, such as asking for restaurant recommendations with photos of the restaurant’s interiors included. At the time of writing, OpenAI’s ChatGPT does not allow image uploads, the chatbot is still entirely text-based, and Anthropic’s chatbot, Claude 2, works similarly.
“Alcohol maven. Evil bacon lover. Wannabe social media geek. Travel guru. Amateur introvert. Pop culture nerd.”