Microsoft has confirmed its investment in the hottest AI company and probably the hottest tech product at the moment, OpenAI.
But why?
On the surface, it doesn’t feel like the two have much in common. But under the microscope (i.e. what you’re about to read), they just might be a match made in heaven.
In this article, we’ll look at key questions about this acquisition.
OpenAI is a research laboratory where several artificial intelligence (AI) applications are trained to mimic human tasks.
Some of these applications and the tasks they mimic include:
The OpenAI research lab is encompassed in both the non-profit, OpenAI Incorporated (OpenAI Inc.) and its for-profit subsidiary, OpenAI Limited Partnership (OpenAI LP).
It was founded by Sam Altman, Greg Brockman, Elon Musk, Wojciech Zaremba, Ilya Sutskever, and John Schulman in San Francisco in 2015.
Its mission is to “ensure that Artificial General Intelligence benefits all humanity”.
There are two goals in this mission that will help you understand how this technology may be deployed in Microsoft Teams moving forward.
Several of OpenAI’s founders have expressed concern for the future of artificial intelligence.
CEO, Sam Altman, has expressed concerns about its use for things like revenge p*rn generation and has even predicted a potential worst-case scenario to be “lights out for all of us”.
So, it’s clear the company founders have a goal of making AI applications that can’t be used maliciously.
Their Alignment Research is geared specifically towards this goal of making sure that they produce Artificial General Intelligence that is aligned with human values and intent for any given task. In other words, they want to ensure their AI doesn’t go rogue.
This is important as you’ll see in OpenAI’s use cases for Microsoft Teams later in this article.
Artificial General Intelligence (AGI) is the ability of an intelligent agent to mimic humans in any intellectual task.
AGI will be the climax of the AI story, and it’s a major plot point in the entire story of computers.
Sam Altman believes that AGI will be attainable within the next decade.
But what do these two goals we’ve just described have to do with Microsoft Teams?
The goal of alignment can inform us of what AI additions to expect in Teams immediately.
Meanwhile, the AGI goal will give us a glimpse of what to expect in the future from a machine that will outperform us at “most economically viable work”.
Microsoft Teams is one of the many products belonging to Microsoft Corporation, which in turn, has been involved in OpenAI over the years.
Besides their $10 billion investment in the AI company, Microsoft has prior investments in 2021 and 2019.
What’s most interesting is the 2019 splurge which involved a $1 billion deal and a partnership that granted Microsoft exclusivity as OpenAI’s cloud provider. The two committed to building new AI supercomputing technologies for the Azure cloud platform.
You see, ChatGPT infamously accrues “eye-watering” costs.
And seeing as OpenAI applications like ChatGPT or DALL·E 2 won’t generate income through the pay-per-click model, it needs monetization.
OpenAI has launched the paid version of ChatGPT at $20/month, but that is not enough to keep the whole operation running.
This is why OpenAI will be collaborating with Microsoft Azure, the world’s second-largest cloud provider, to help incorporate their AI models into other companies’ software.
An example is Notion AI which is powered by the GPT-3 language model.
Consider this business model “AI-as-a-service” (AIaaS).
And Microsoft’s platforms, like Microsoft Teams, have to be among the first to incorporate OpenAI technology.
On Monday, January 23, 2023, Microsoft confirmed the rumors of the “multiyear, multibillion-dollar” partnership with OpenAI.
It confirmed deployment of OpenAI’s models to bring “new AI-powered experiences” across consumer and enterprise Microsoft products.
Perhaps the most anticipated was the introduction of ChatGPT-like features to Bing.
And on the 7th of February, 2023, Microsoft announced the new and improved Bing as expected.
In February, Microsoft Teams Vice President, Nicole Herkowitz, announced Microsoft Teams Premium.
Teams Premium is a new offering that will “cut costs and add AI-powered productivity”.
The premise is that organizations no longer have to pay for add-ons that enhance meeting capabilities. Instead, Teams Premium will provide these capabilities at just $10/user/month.
In the blog post, Herkowitz listed two key AI additions coming to Microsoft Teams:
This is the sum of new features that answer the question “what did I miss?” after meetings.
These intelligent recap features include:
At the moment, only AI-generated chapters and personalized time markers are available.
By Q2 2023, Teams Premium is expected to include speaker timeline markers, AI-generated notes, AI-generated tasks, and personalized time markers.
Teams Premium will also enable AI-powered translations for meeting participants in up to 40 languages.
All attendees in a meeting can translate captions to their own language if the meeting organizer has Teams Premium.
These are other applications of AI that would be exciting if deployed in Microsoft Teams.
The Adobe team is working on Project Shasta. It’s an AI-powered recording tool that lets you edit your audio using words instead of waveforms and enhances your audio recording with the click of a single button.
No special editing skills, no special microphones, just AI-powered enhancement.
Maxine is Nvidia’s own platform for deploying AI, and one of its most impressive features is “Eye Contact”.
This feature enables users to keep their gaze on the camera during a recording or a meeting, in real-time.
This feature may also be useful for “content creators seeking to record themselves while reading their notes or a script”.
Teams could integrate the chatbot for organizations to get quick answers, ideas, and more without having to leave the Teams app.
Less context switching and more debugging, code creating, or whatever is you’re working on.
If we entertain the more sci-fi use cases, there’s also the scenario of training GPT-3 in Teams with users' texts and conversations. Your personal chatbot could give smart replies on your behalf for the most basic things.
But it’s not really “sci-fi”, is it? This is what chatbots do before relaying to a real person for a more “human” touch.
GPT-3 in Teams could take this “economically viable work” and put it in your hands.
We asked ChatGPT about Microsoft and OpenAI. This is how that conversation went.
Prompt: What do you think OpenAI's role will be in Microsoft Teams?
ChatGPT: As an AI language model created by OpenAI, I don't have opinions or emotions, but I can tell you that Microsoft has recently announced plans to integrate OpenAI's language model GPT-3 into Microsoft Teams, the company's popular collaboration and communication platform.
The integration of GPT-3 into Microsoft Teams will likely provide users with advanced language capabilities, such as improved spelling and grammar suggestions, as well as conversation suggestions and smart replies.
This could help users to save time and be more productive by quickly generating content, responding to messages, and creating new documents.
The integration of GPT-3 into Microsoft Teams is part of Microsoft's broader strategy to integrate AI and machine learning into its products and services, with the goal of improving the overall user experience and making people more productive.
Here’s what Microsoft MVPs are saying you should expect from OpenAI in Microsoft Teams.
Solutions Architect at Symity, Lee Ford, described a potential use case for DALL·E 2 in generating images within Microsoft Teams conversations.
Microsoft Developer Advocate, Garry Trinder, has built a bot to turn text into images using OpenAI and Microsoft Teams.
Tim Russell, a Partner Technical Architect at Microsoft, believes, “OpenAI in Teams can be a gamechanger if it can be implemented for conversational searching of all Microsoft 365 artifacts.”
Tom Arbuthnot, Founder of Empowering.Cloud suggests “the potential of GPT-3 could be useful for smarter replies, meeting summaries, pull out actions, and chasing people on actions.” He also cites potential use cases in sentiment analysis and coaching users on how to use Teams more effectively.
But your AI constantly popping up to tell you what to do brings back memories of one of Microsoft's previous failures at adding intelligent software in Teams.
Tom Morgan, Microsoft MVP and Development Lead at Empowering.Cloud, mentions that one of the incredible yet mostly-ignored achievements of OpenAI has been to abstract all the complexity.
“It’s ridiculously easy to use their APIs, and with one call you can generate an image, produce a news article, write code, or a thousand other things. OpenAI has abstracted years of research, computational algorithms that would make your head bleed and hours of painstaking human training into a single line of code for those consuming it.”
Tom says that’s amazing and commendable, but it’s not the end of the story.
“To take OpenAI to the next level, we need to apply it to actual problems that real people have. Some of this will come from Microsoft but I hope much of it will come from the community of Microsoft Teams developers. In fact, I’m going to say that citizen developers using low-code tooling are better placed than professional developers here because they tend to understand their problem domain better.”
You see, Microsoft has tried its hands at intelligent assistants in the past.
Office Assistants like Clippy were largely failures. But the introduction of OpenAI in Teams has left many wondering if Clippy and co. will be making a comeback.
There’s also the curious case of Cortana in Teams, which many people also didn’t like.
With Clippy, there was so much interruption, and with Cortana one often forgets it exists (don’t you?). So is OpenAI in Microsoft Teams going to be a new type of disaster?
Who knows?
Mass adoption shows it’s prime for success. Nevertheless, Microsoft Teams users have to keep their fingers crossed