new york
CNN
—
In just a few months you will be able to ask be With a virtual assistant, you can transcribe meeting notes during a work call, summarize long email threads to quickly create suggested replies, quickly create a specific chart in Excel, or create a Word document in seconds. to a PowerPoint presentation.
It’s only on Microsoft’s 365 platform.
Over the past week, the rapidly evolving landscape of artificial intelligence has seemed to take off again. Microsoft and Google each announced new AI-powered features for their signature productivity tools, and OpenAI introduced the next-generation version of the technology behind its viral chatbot tool, ChatGPT.
The AI tools that have been working in the background of many services for so long have suddenly become more powerful and visible in an ever-wider range of workplace tools.
For example, Google’s new feature promises to help you “brainstorm” and “proofread” your work written in documents. in the meantime, If your workplace uses the popular chat platform Slack, you might be able to ask its ChatGPT tool to converse with your co-workers, write and reply to new messages, or summarize conversations in channels.
OpenAI, Microsoft, and Google are at the forefront of this trend, but they’re not alone. IBM, Amazon, Baidu and Tencent are also working on similar technologies. Many startups are also developing AI lighting assistants and image generators.
The tech company’s pitch is clear. AI increases productivity and eliminates tedious tasks. As Microsoft CEO Satya Nadella said in a presentation on Thursday, “We believe this next-generation AI will usher in a new wave of productivity improvements. It takes the drudgery out of work and makes us rediscover the joy of creation.” ”
But the sheer number of new options coming to market is dizzying, and like many other things in the tech industry over the last decade, Failing to live up to the hype can have unintended consequences, such as enabling cheating or eliminating the need for certain roles (although that is the intent of some recruiters). is also possible).
Even the promise of increased productivity is vague. For example, the rise of AI-generated email can improve sender productivity, but reduce the productivity of recipients who are bombarded with computer-generated messages that are longer than necessary. And of course, just because everyone has the option to use chatbots to communicate with their colleagues doesn’t mean everyone will choose to do so.
Forrester analyst Rowan Curran said that integrating this technology into the foundation of the productivity software most of us use every day will have a profound impact on how we work. would,’ he said. “But that change won’t sweep everyone and everyone tomorrow. It will take time to learn how to make the most of these features to enhance and adjust existing workflows.”
If you’ve ever used the autocomplete option when typing an email or sending a message, you’ve already experienced how AI can speed up your tasks. But the new tool promises to go far beyond that.
A new wave of AI product launches began about four months ago when OpenAI released a limited version of ChatGPT., It surprises users by generating human-like responses to user prompts, passing prestigious university exams, and writing persuasive essays on various topics.
Since then, the technology, in which Microsoft invested “billions of dollars” earlier this year, has only gotten better. Earlier this week, OpenAI announced his GPT-4, a more powerful version of the technology behind ChatGPT. This promises to blow previous iterations out of the water.
In early tests and in-house demos, GPT-4 was used to draft lawsuits, build working websites from hand-drawn sketches, and create iconic apps like Pong, Tetris, and Snake with little coding experience. I reproduced the game.
GPT-4 is a large-scale language model trained on vast amounts of online data to generate responses to user prompts.
This is the same technology that powers two new Microsoft features. A “co-pilot” that helps you edit, summarize, create and compare documents across platforms, and a business her chat that is basically an agent who accompanies users on their work and trials. To understand and make sense of your Microsoft 365 data.
For example, agents know what’s in your email, what’s in your calendar for the day, what documents you’re working on, presentations you’re working on, people in meetings, and so on. The company says the chats are hosted on the Teams platform. Users can then ask Business Chat to perform tasks such as summarizing all documents across the platform for a particular project to create a status report, or email to send updates to the team. You can create drafts of
Karan said The degree to which these AI-powered tools transform your work depends on your application. For example, word processing applications help create outlines and drafts, slideshow programs help speed up the design and content creation process, and spreadsheet apps allow more people to work with data and make data-driven decisions. help you do that. He believes the latter will have the most significant impact on the workplace in both the short and long term.
Discussions about how these technologies will impact work “should focus on the task of the job, not the job as a whole,” he said.
OpenAI’s GPT-4 update promises to fix some of its biggest challenges, including the potential for perpetuating biases, sometimes getting the facts wrong, and responding in aggressive ways. Problems can still occur. Some of these issues affect the workplace, especially in interactions with others.
Arijit Sengupta, CEO and founder of AI solutions company Aible, said the problem with large language models is that they often try to please their users and accept assumptions about what they say.
“When people start gossiping about something, we accept it as the norm and start generating content. [related to that]Sengupta said, adding that interpersonal issues can escalate into workplace bullying.
OpenAI CEO Sam Altman tweeted earlier this week The technology behind these systems “still has flaws, still has limitations, and seems more impressive when first used than after extended use,” the authors write. I’m here. In a blog post, the company reiterated, “Extreme care must be taken when using language model output, especially in high-stakes contexts.”
Gartner Research analyst Arun Chandrasekaran says organizations need to educate users about the benefits and limitations of these solutions.
“Blind trust in these solutions is as dangerous as complete disbelief in their effectiveness,” Chandrasekaran said. “Generative AI solutions can also fabricate facts or present inaccurate information. Organizations should be prepared to mitigate this negative impact.”
At the same time, many of these applications are not up to date (data from GPT-4 used for training) end around September 2021). You are responsible for doing everything from double-checking accuracy to changing the language to reflect your desired tone. Also, gaining buy-in and support across the workplace is critical to the adoption of the tool.
“Training, education and organizational change management are very important to ensure that employees are upholding the initiative and the tool is being used as intended,” Chandrasekaran said.
