Yes, the following was actually written by GPT-4, as integrated into the Microsoft Edge “Discover” panel. This paragraph is the only one that I wrote by hand, and I have to say… I have mixed feelings. I’ve left the computer output exactly as it was generated, including a punctuation error. Furthermore, I fact checked the ability to generate images with Microsoft Bing since that is not currently a feature, nor did I see it in the blog post referenced by the AI writer. As it turns out, Bing is launching an AI powered tool called Image Creator, but I don’t yet have access to it.
Microsoft adopts GPT-4 into Bing and Edge
Microsoft has announced that it has integrated OpenAI’s latest language model, GPT-4, into its Bing search engine and Edge web browser. This is a major milestone for both companies, as GPT-4 is the most advanced artificial intelligence system of its kind, capable of generating natural language texts and images from any input.
GPT-4 was officially unveiled by OpenAI on March 14th, 2023, after months of speculation and rumors. The AI model has 1 trillion parameters, 10 times more than its predecessor GPT-3, which was already considered a breakthrough in natural language processing. GPT-4 can not only generate coherent and diverse texts on any topic, but also produce realistic images from text descriptions or captions.
However, what many people did not know was that GPT-4 was already powering Bing since February 2023, as part of a secret collaboration between Microsoft and OpenAI. Microsoft had access to an early version of GPT-4 through its exclusive license agreement with OpenAI, which gave it the right to use the AI model for its own products and services. Microsoft decided to test GPT-4 on Bing as a way to improve its search quality and user experience.
According to Microsoft’s blog post, Bing users can now enjoy several benefits from using GPT-4:
- More relevant and comprehensive search results: GPT-4 can understand complex queries better than traditional algorithms, and provide more accurate and diverse answers from multiple sources.
- More engaging and interactive search experience: GPT-4 can generate natural language responses to user questions or comments, as well as suggest related topics or queries for further exploration.
- More visual and creative search options: GPT-4 can create images from text inputs or modify existing images based on user requests. For example, users can ask Bing to show them “a cat wearing a hat” or “a sunset with purple clouds” and get realistic images generated by GPT-4.
Microsoft has also integrated GPT-4 into its Edge web browser, which now features a new sidebar that allows users to chat with Bing’s AI chatbot. The chatbot can answer user questions about any topic, provide useful information or tips, or even generate fun content such as jokes or stories. Users can also ask the chatbot to show them images created by GPT-4 based on their inputs.
Microsoft claims that using GPT-4 in Bing and Edge will enhance user satisfaction and loyalty, as well as attract new users who are curious about the AI capabilities. Microsoft also says that it will continue to work with OpenAI to improve GPT-4’s performance and safety, as well as explore new ways to apply it across its products and services.
GPT-4 is currently available only through OpenAI’s API, which requires developers to join a waitlist and pay a fee based on usage. Microsoft is one of the few companies that have direct access to GPT-4 through its license agreement with OpenAI. However, OpenAI says that it plans to make GPT-4 more accessible in the future, as part of its mission to ensure that artificial intelligence benefits humanity as a whole.
