In just a few months, you will be able to ask A a virtual assistant to transcribe meeting notes during a work call, summarize long email streams to quickly put together suggested responses, quickly create a specific chart in Excel, and turn a Word document into a PowerPoint presentation in seconds.
And that’s just on Microsoft 365 platforms.
Over the past week, the fast-paced artificial intelligence landscape seemed to take the lead again. Microsoft and Google have unveiled new AI-based features for their proprietary productivity tools, and OpenAI has unveiled a next-generation version of the technology behind its viral chatbot tool ChatGPT.
Suddenly, AI tools that have been running in the background for a long time across many services are now more powerful and more visible across a wide and growing range of workplace tools.
New features of Google, for example, promise for brainstorming and proofreading written papers in Documents. Meanwhile, If your workplace uses the popular chat platform Slack, you can use its ChatGPT tool to communicate with co-workers on your behalf, perhaps by asking them to write and reply to new messages and summarize conversations in channels.
OpenAI, Microsoft and Google are at the forefront of this trend, but they are not alone. IBM, Amazon, Baidu and Tencent are working on similar technologies. A long list of startups are also developing AI writing assistants and image generators.
The position of technology companies is clear: AI can make you more productive and save you from routine work. As Microsoft CEO Satya Nadella said during Thursday’s presentation, “We believe the next generation of AI will usher in a new wave of productivity growth: powerful assistants designed to take us out of the grind of our daily tasks and at work, giving us the power to rediscover for yourself joy. creativity.”
But the sheer number of new options hitting the market is dizzying and, like so much else in the tech industry over the past decade, raises questions about whether they really are. justify the hype or cause unforeseen consequences, including the possibility of fraud and the elimination of the need for certain roles (although this may be the intention of some followers).
Even the promise of greater performance is unclear. For example, an increase in the number of AI-generated emails can improve sender productivity, but degrade it for recipients inundated with longer-than-necessary computer-generated messages. And of course, just because everyone has the ability to use a chatbot to communicate with colleagues doesn’t mean everyone will choose to do so.
Integrating this technology “into the core pieces of productivity software that most of us use every day will have a significant impact on how we work,” said Rowan Curran, an analyst at Forrester. “But this change won’t affect everything and everyone tomorrow – learning how to best use these opportunities to improve and adjust our existing workflows will take time.”
Anyone who has ever used the autocomplete option when typing an email or sending a message has already experienced how AI can speed up tasks. But new tools promise to go far beyond that.
A new wave of AI product launches began almost four months ago when OpenAI released a limited version of ChatGPT., wow users by generating human responses to user queries, taking exams at prestigious universities, and writing compelling essays on a range of topics.
Since then, the technology, in which Microsoft made a “multi-billion dollar” investment earlier this year, has only gotten better. Earlier this week, OpenAI introduced GPT-4, a more powerful version of the technology behind ChatGPT that promises to outperform previous iterations.
In the company’s first tests and demos, the GPT-4 was used to draft lawsuits, build a working website from a hand-drawn sketch, and recreate iconic games like Pong, Tetris, or Snake with very little or no programming experience.
GPT-4 is a large language model that has been trained on massive amounts of online data to generate responses to user queries.
It’s the same technology that powers two of Microsoft’s new features: “Co-pilot” to help you edit, summarize, create and compare documents across its platforms, and Business Chat, an agent that essentially rides with the user. when he is working and trying to understand and make sense of his Microsoft 365 data.
The agent will know, for example, what’s in the user’s email and in his calendar for the day, as well as the documents he worked on, the presentations he made, the people he met, and according to the company, the chats happen on their platform. commands. Users can then ask Business Chat to complete tasks such as writing a status report summarizing all documents across platforms for a particular project, and then compose an email that can be sent to their team with an update.
Curran just said How much these AI-based tools will change the way you work depends on the application. For example, a word processing application can help create plans and drafts, a slideshow maker can help speed up the design and content creation process, and a spreadsheet application should help more users interact and make data-driven decisions. The latter, in his opinion, will have the most significant impact on the workplace in both the short and long term.
Discussions about how these technologies will affect jobs, he says, “should be focused on work tasks, not jobs in general.”
While the OpenAI GPT-4 update promises to fix some of its biggest problems – from its ability to perpetuate prejudice to sometimes factually incorrect and aggressive reactions – there is still room for some of these issues find their way into the workplace, especially when it comes to interacting with other people.
Arijit Sengupta, CEO and founder of artificial intelligence solutions company Aible, said the problem with any large language model is that it tries to please the user and usually accepts the premises of the user’s statements.
“If people start gossiping about something, they will accept it as the norm, and then they will start generating content. [related to that]Sengupta said, adding that it can escalate interpersonal issues and turn into office bullying.
In a tweet earlier this week, OpenAI CEO Sam Altman wrote that the technology behind these systems is “still imperfect, still limited, and still feels more impressive the first time you use it than after you’ve spent more time with it.” The company reiterated in a blog post that “great care should be taken when using language model output, especially in high-stakes contexts.”
Arun Chandrasekaran, an analyst at Gartner Research, said organizations need to educate their users on what these solutions are good at and what their limitations are.
“Blind trust in these solutions is just as dangerous as a complete lack of faith in their effectiveness,” Chandrasekaran said. “Generative AI solutions can also falsify facts or provide inaccurate information from time to time, and organizations must be prepared to mitigate these negative impacts.”
At the same time, many of these applications are outdated (GPT-4 data, on which he was trained ends around September 2021). Users will have to do everything from double checking for accuracy to changing the language to reflect the desired tone. It will also be important to get approval and support from the workplace in order for the tools to start working.
“Training, learning and organizational change management are very important to ensure that employees support efforts and tools are used as intended,” said Chandrasekaran.