As AI assistants such as ChatGPT continue to cause a sensation, a survey by Microsoft of over 3,000 people has found that the vast majority of workers want more help from artificial intelligence to automate mundane daily tasks.
Microsoft’s WorkLabs survey attempts to size up worker and employee expectations at a time of economic uncertainty, rapid technological change, and emerging expectations about hybrid work. It surveyed 2,700 workers and 1,800 business decision makers in the US, UK, and Japan.
The research found that 89% of people feel “more fulfilled” when they have access to AI tools “because they can spend time on work that truly matters”. But 30% of employees surveyed said they don’t have access to AI tools. Of those with access to AI tools, 54% said AI features help with problem-solving tasks, and 89% hope they can use AI for more tasks and activities.
The findings fit with Microsoft’s own strategy for AI and tools such as ChatGPT. The tech giant recently made a big investment in OpenAI, the company behind ChatGPT, and has said it will deploy OpenAI’s models across its consumer and enterprise products, and “introduce new categories of digital experiences built on OpenAI’s technology”.
The findings come as educators decide whether and how to regulate ChatGPT and other AIs amid fears that students could use the technology to write home assignments and cheat in exams. Widespread use could undermine the most efficient and trusted methods for testing a student’s knowledge. New York City Department of Education banned ChatGPT for both staff and students due to concerns over the accuracy of the technology’s output and fears it may harm student learning.
But a professor at Wharton, University of Pennsylvania found ChatGPT could formulate clever questions suitable for an MBA exam and liked its potential to reduce his need to reformulate similar questions every year. ChatGPT also passed the equivalent of a typical MBA exam.
Researchers have also found that ChatGPT is adept at fixing software bugs, but have questioned whether the mental cost to verify ChatGPT outweighs the advantages it brings. StackOverflow banned ChatGPT because its plausible-looking but wrong answers overwhelmed site content moderators.
January 27, 2023
Written by Liam Tung