Microsoft Launches Copilot, Google Adds Generative AI to Workspace, More News
GPT-4 is here. In a statement this week, San Francisco-based OpenAI announced the release of the new, multimodal version of the technology which can generate content based on images as well as text.
The text-input feature is available to ChatGPT Plus subscribers — the technology's first paid plan — and to software developers via a waitlist, while the image-input ability remains in preview.
However, despite the flurry of headlines and reactions following the release, the company has continued to downplay expectations. In various statements, the company has reiterated that while GPT-4 exhibits human-level performance in a number of professional and academic benchmarks, it is less capable than humans in many real-world scenarios. As recently as last December, OpenAI CEO Sam Altman admitted that ChatGPT creates "a misleading impression of greatness" and that it "hallucinates" facts. He urged people to use "great care" when using language model outputs, particularly in "high-stakes contexts."
In a tweet coinciding with the release, Altman called the latest release "still flawed, still limited."
here is GPT-4, our most capable and aligned model yet. it is available today in our API (with a waitlist) and in ChatGPT+.https://t.co/2ZFC36xqAJ
it is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it.— Sam Altman (@sama) March 14, 2023
However, the improvements touted in the release show an exponential leap forward from GPT-3.5 and GPT-4. For example, the company noted in a statement that GPT-4 scored in the top 10% of test takers in a simulated bar exam. GPT-3.5’s score was closer to the bottom 10%.
OpenAI outlined some — thought definitely not all — of the evolution of the technology in the release. Over the past two years, the company has rebuilt its deep learning stack from scratch using Azure, as well as building a supercomputer from the ground up to manage its workload. A year ago, the company explained, it developed GPT-3.5 as a first “test run” of the system. “We found and fixed some bugs and improved our theoretical foundations. As a result, our GPT-4 training run was (for us at least!) unprecedentedly stable, becoming our first large model whose training performance we were able to accurately predict ahead of time,” the statement read.
The difference between GPT-3.5 and GPT-4 is substantial, which become clear as tasks grow more complex. In these situations, GPT-4 is more reliable, creative and able to handle much more nuanced instructions than GPT-3.5.
OpenAI has already partnered with a number of companies to integrate GPT-4 into their products, including Duolingo, Stripe and Khan Academy. More importantly, Microsoft confirmed it has been running its new Bing search engine on GPT-4.
“We are happy to confirm that the new Bing is running on GPT-4, which we’ve customized for search. If you’ve used the new Bing preview at any time in the last five weeks, you’ve already experienced an early version of this powerful model,” Microsoft’s statement read.
Microsoft Announces Copilot, Bringing GPT-4 to Microsoft 365
As predicted, Microsoft wasted no time pulling GPT into Microsoft 365. Two days after the release of GPT-4, Microsoft announced the new Copilot.
The Redmond, Wash.-based company announced Copilot during a live-streamed presentation on March 16. The company billed Copilot as "the most powerful productivity tool on the planet." In short, Copilot works in tandem with other Microsoft apps throughout the Microsoft 365 platform to help people write documents, develop PowerPoint presentations, apply PivotTables in Excel and more. It will also be used for what Jared Spataro, corporate vice president of modern work and business applications described as "an entirely new chat experience" called Business Chat.
Business Chat, he explained in a blog following the event, applies large language models (LLMs) to data in the Microsoft Graph and the Microsoft 365 apps — including those running calendar, emails, chats, documents, meetings and contacts — to provide better insights and more accurate answers to questions about an organization's data.
“You can give it natural language prompts like, 'Tell my team how we updated the product strategy,' and it will generate a status update based on the morning’s meetings, emails and chat threads," Spataro wrote.
Spataro explained the overall result would be to make workers better at what they are good at. “It lets you quickly master what you’ve yet to learn. The average person uses only a handful of commands — such as 'animate a slide' or 'insert a table' — from the thousands available across Microsoft 365. Now, all that rich functionality is unlocked using just natural language. And this is only the beginning," he wrote.
It's unclear how such an enormous step forward for Microsoft 365 actually works. Microsoft is currently testing Copilot with 20 customers, with promises to expand into previews in the coming months.
The large language models driving Copilot have real-time access to Microsoft Graph, which maps an organizations' content and context to provide more accurate answers to queries based on that content. Microsoft 365 and Future of Work general manager Colette Stallbaumer wrote Copilot abides by the security protocols and privacy settings of an organization and also follows the company's Responsible AI Standards (this coming the same week the company fired the remaining members of its AI ethics team).
In mentioning the previews, the company assured it would communicate with customers and specifically, their IT admins, to best prepare for the introduction of Copilot.
What comes next is anyone’s guess but what we do know is Copilot will be brought into Word, Excel, PowerPoint, Outlook, Teams, Viva (with the announcements specifically calling out Viva Engage), Power Platform and other apps.
Pricing for Copilot wasn't included in the announcement, but it should be coming soon, as well as implications for existing licenses.
Google Brings Generative AI to Workspace
Microsoft’s announcement comes just days after Google announced similar generative AI features for Google Workspace, including AI-assisted text generation in Gmail, Docs and more. The announcement came with promises of more to come.
Learning Opportunities
Google is describing the development as the biggest thing to happen to its productivity suite since the addition of collaboration features.
According to a post about the new updates, people working with Gmail and Google Docs can produce draft emails and documents by writing a prompt with the topic in mind. The writer can then edit the draft before sending or sharing it.
The blog claimed more than three billion people are already using AI in Google Workspace through its Smart Compose and auto-generated summaries in Google Docs.
Using what it describes as “trusted testers,” Google says as well as the upcoming Docs and Gmail features, the new AI additions will also:
- Enable the creation of auto-generated images for Slides as well as the creation of videos and images.
- Add a process to add contextual data categorization in Sheets built on raw data
- Build new capture notes and background to Meet
Enable Chat workflows.
Clearly this is more than the usual upgrades and add-ons that Google offers on a regular basis.
An accompanying post from Thomas Kurian CEO, Google Cloud, stated that Google (and presumably other productivity vendors) have reached a pivotal point in the use of workplace technology.
“Breakthroughs in generative AI are fundamentally changing how people interact with technology — and at Google, we’ve been responsibly developing large language models so we can safely bring them to our products,” he wrote.
Much of this is quite forward thinking and apart from the Docs and Gmail additions it is unclear when the other additions will land and how they will work.
Kurian also outlined the release of a new API through which developers can access Google’s foundational AI models, starting with its PaLM large language model for building language-oriented apps.
This is likely to be just the beginning. While we are used Microsoft and Google banging against each other in the productivity space, these announcements will push the competition to an entirely new level.
Omneky Releases New Chatbot with ChatGPT API
Finally this week, San Francisco-based Omneky announced the launch of its Creative Assistant tool, a chat-based interface based on OpenAI’s ChatGPT API (GPT-3.5 Turbo). The Creative Assistant tool allows organizations to develop a creative brief, complete with AI generated visual storyboard and mood boards, further expanding Omneky’s generative digital platform.
The AI-powered chatbot prompts users for preliminary information about the intended product or campaign as well as tone. Omneky also recently launched the Custom AI-Generated Product Imagery tool, which has two functions: producing rapid iterations of product images based on catalog photos and generating custom image backgrounds.
Founded in 2018, Omnesky has raised $10 million to date in two funding rounds.
Have a tip to share with our editorial team? Drop us a line:
About the Author
David is a full-time journalist based in Ireland. A partisan of ‘green’ living and conservation, he is particularly interested in information management and how enterprise content management, analytics, big data and cloud computing impact on it.
Featured Research
Research Report
Read Now
eBook
Read Now
Guide
Read Now
Guide
Read Now
Research Report
Read Now
Buyer's Guide
Read Now
Infographic
Read Now
White Paper
Read Now
Related Stories
Smarter Experiences Win — Learn from Today's Leaders
