For First Time, Microsoft Integrating GPT-3 Into Its Software
MICROSOFT BUILD 2021 – Eight months after licensing the GPT-3 natural language AI model from OpenAI last September, Microsoft is integrating the language generator into its Microsoft Power Apps software to make it easier for enterprise workers to build no-code applications.
The news, which was announced May 25 (Tuesday) at the virtual Microsoft Build 2021 conference, marks the first time that Microsoft is using OpenAI’s GPT-3 model within one of its own applications aimed at customers.
Running on Microsoft Azure and powered by Azure Machine Learning, the use of GPT-3 within Microsoft Power Apps shows how GPT-3 can solve real-world business needs on an enterprise scale, according to a May 25 post on the Microsoft Blog.
Once GPT-3 is integrated with Microsoft Power Apps, non-technical employees will be able to build a no-code Power Apps application by entering conversational language and then have it automatically transformed into the needed code using GPT-3, according to Microsoft. Power Apps code is written in the open source Power Fx language, which is inspired by Microsoft Excel. Though Power Fx is easier to use than traditional coding languages, there can still be a steep learning curve to create complex data queries, which is why integrating GPT-3 was seen as a good option to make things even easier for so-called “citizen developers” who need to build apps but lack the proper training and skills. It can also be used by professional developers and anyone in-between.
“For instance, the new AI-powered features will allow an employee building an e-commerce app to describe a programming goal using conversational language like ‘find products where the name starts with ‘kids,’” the post states. A tuned GPT-3 model would then offer choices to automatically transform the command into a Microsoft Power Fx formula, such as “Filter(‘BC Orders’ Left(‘Product Name’,4)=”Kids,” according to the post.
“GPT-3 is the most powerful natural language processing model that we have in the market, so for us to be able to use it to help our customers is tremendous,” Bryony Wolf, Power Apps product marketing manager, said in the post. “This is really the first time you’re seeing in a mainstream consumer product the ability for customers to have their natural language transformed into code.”
GPT-3, which is built by OpenAI, an independent AI research and deployment company, is a massive natural language model that runs exclusively on Microsoft Azure. The use of GPT-3 in Power Apps is only the latest collaboration between the two companies. In 2019 the companies paired to develop and expand new joint Azure AI supercomputing capabilities, while in 2020 they collaborated on building an AI supercomputer that was hosted in Azure to train OpenAI’s models. In addition, Microsoft invested $1 billion in July of 2019 to support the company's AI work.
The new GPT-3 features are expected to be available in preview in English throughout North America by the end of June, according to Microsoft.
The use of GPT-3 in other Microsoft products is also being eyed, the company said.
“We’re finding ways to bring it into Azure and our mainstream products,” Eric Boyd, Microsoft’s VP for Azure AI, said in the post. “We think there are a whole bunch more things that GPT-3 is capable of doing. It’s a foundational new technology that lights up a ton of new possibilities, and this is sort of that first light coming into production.”
Karl Freund, founder and principal analyst with Cambrian AI Research, told EnterpriseAI that the pairing of GPT-3 with Microsoft Power Apps is “quite impressive.”
“When GPT-3 was first announced, OpenAI and Microsoft claimed it would become a big deal, that it could open the door for many new applications built with little or no programming skills,” said Freund. “Now we can see the first real use cases emerge on Microsoft Azure and I believe we will see many more in the future. AI is changing how computers are fundamentally used and programmed. This is super exciting stuff.”
GPT-3, which stands for Generative Pre-trained Transformer 3, is an autoregressive language model with 175 billion parameters, which OpenAI claims is ten times more than any previous non-sparse language model. This allows GPT-3 to achieve remarkable results in many translation, question-answer and text generation tasks with no fine-tuning using only a small amount of training data. There were two earlier versions of GPT – GPT-2 and the original GPT.
Microsoft exclusively licensed GPT-3 from OpenAI in September 2020, extending the relationship between the two companies. The licensing agreement covers the use of GPT-3 for all Microsoft products and services. Other users can still use the GPT-3 model through OpenAI’s API, according to OpenAI.
The Microsoft Build 2021 conference continues through May 27.