Microsoft's campaign to let enterprise IT users easily integrate ChatGPT generative AI technology into business, consumer and developer applications pushed forward on Tuesday as the cloud giant sought to dominate the AI arms race.
The vendor introduced a slew of updates to ChatGPT and the Bing browser, a new plugin platform, and partnerships with key tech players, including AI hardware/software vendor Nvidia.
The product introductions and other developments came at Microsoft Build, the vendor's developer conference held in-person in Seattle and streamed to a virtual audience.
"This is about really creating that opportunity for developers to reach all users across all of these surface areas," Microsoft CEO Satya Nadella said, referring to the vendor's myriad developer and business applications.
ChatGPT and Bing
One way the tech giant is helping developers reach a wider audience is by making Bing Search part of OpenAI's ChatGPT.
The change will not only provide ChatGPT users with Microsoft's search engine but also make ChatGPT, which previously was trained on data up to 2021 only, up to date. Users of the paid service can ask ChatGPT real-time queries immediately, remedying what has been a major gap for ChatGPT.
Users of the free version users will gain access to the up-to-date ChatGPT later with a plugin, Microsoft said. The vendor did not set a specific date.
The move to make Bing Search part of ChatGPT is logical, especially since large language models such as ChatGPT are much more helpful when they can connect with the world in real time, said Forrester Research analyst Rowan Curran.
The closer connection between Bing Search -- a longtime competitor to Microsoft archrival Google's Search -- also makes sense as Microsoft brings many new plugins to its ecosystem, Curran added.
Microsoft revealed that it has adopted the same open plugin standard as its partner in generative AI OpenAI uses.
Developers now have access to more than 50 plugins on a new platform that let users plug into consumer and business applications, including ChatGPT, Bing, Microsoft 365 Copilot and Dynamics 365 Copilot.
Developers can also create, test and deploy their plugins.
Rowan CurranAnalyst, Forrester Research
"The plugins are a 100% necessary step to make these large language models useful as they're being deployed," Curran said. "These plugins are basically allowing these models to … not just be brains in a box. They're allowing them to have hands and feet and other ways of interacting with the world."
Instead of just asking questions of the LLMs, developers and other users, through the plugins, can use them to interact with and execute practical actions in the real world, such as setting up travel or shopping.
The plugin strategy is also an effective strategy for the tech giant to make AI available across its portfolio, especially through its Copilot applications, said Karl Freund, founder and analyst at Cambrian AI.
Microsoft defines Copilots as applications that use AI and LLMs to help users perform cognitive tasks.
By using plugins to connect to different Copilot applications, "Microsoft is becoming the leading cloud provider for AI applications," Freund said.
While Google will continue to be a worthy competitor, it doesn't possess the Copilot capabilities that Microsoft wields across a broad range of applications, he added.
Besides building a bridge through plugins to its Copilot applications, Microsoft is forging key new strategic partnerships.
One of them is an alliance with Nvidia, one of the most dynamic independent players in AI.
Microsoft's Azure Machine Learning service will now be integrated with Nvidia AI Enterprise software. The integration will help Azure customers build, deploy and manage customized applications using 100 Nvidia AI frameworks and tools while supported by Nvidia's AI platform.
"This is natural for Microsoft Azure," Freund said.
This partnership provides Microsoft users with the same benefit that Nvidia Enterprise AI -- a platform for building and customizing AI products -- offers, he said.
Enterprises that want to build domain-specific version of ChatGPT can use Nemo, Nvidia's cloud-native framework for building LLMs, to do that.
Despite Microsoft's close relationship with OpenAI, integrating with Nvidia's Enterprise AI fills some hardware and software gaps that software vendor and research organization OpenAI can't.
"Enterprises need to find a path to identify the best AI opportunities from their proprietary data as well as build an AI strategy that handles everything from training to model development to governance," said Daniel Newman, an analyst at Futurum Research.
"Nvidia offers a more complete set of tools for enterprises to build around and work in concert with ChatGPT and other open source LLMs," he continued.
Meanwhile, "Microsoft needs Nvidia chips on their Azure platform, which is primarily where they're going with this to make sure that their OpenAI and Copilot all work successfully," said Jack Gold, an analyst at J. Gold Associates.
Although Nvidia works with multiple cloud providers, including Google, no conflicts of interest exist. Nvidia positions itself as an AI "arms provider" that builds and deploys "arms" for cloud providers to use, Freund said.
Microsoft also revealed a new responsible AI initiative. The cloud provider's new Azure AI Content Safety service aims to help developers create safe online environments with models that detect inappropriate content in images and text. Content Safety is now in preview.
Esther Ajao is a news writer covering artificial intelligence software and systems.