Meet the ‘super users’ who tap AI to get ahead at workWorkers say using AI tools like ChatGPT every day supercharges their efforts and saves them time on the job.
https://www.washingtonpost.com/technology/2024/10/28/ai-work-superusers/
Lisa Ross is a little techie, a tad nerdy, and a bit over analytical. So getting to use artificial intelligence for work couldn’t have been a more tantalizing prize.
Ross, who has attention-deficit/hyperactivity disorder (ADHD) and uses they/them pronouns, says they’ve more than doubled their productivity by using AI for almost everything — to create content using specific frameworks, prepare for tough conversations, help with project and strategic planning, and determine impact. The vice president at Canada-based corporate training company Avenue even uses AI to help create custom bots for future tasks. They’ve already built 40, becoming known as the “AI guy” who can help colleagues use it, too.
A growing number of workers are becoming “super users” of AI. They turn to the tech daily to learn skills, analyze large sets of data, review job candidates and even program other bots to help them with repetitive tasks like building online courses. Workers say AI tools — such as OpenAI’s ChatGPT, Microsoft Copilot and other chatbots powered by large language models — help them boost efficiency and confidence at work, and reclaim hours of time. But those familiar with the technology say they also worry about privacy, inaccuracies, the loss of skills and even the potential of job replacement in the future.
Adoption of AI at work is still relatively nascent. About 67 percent of workers say they never use AI for their jobs compared to 4 percent who say they use it daily, according to a recent survey by Gallup. But those who use it, mostly white-collar workers, see benefits in productivity, efficiency, creativity and innovation, the survey said.
“I’m completely obsessed,” said Ross, 48. “There’s so much to learn, I’m living in this space, and I just love it.”
For some workers, AI not only saves time, it helps them develop skills. Ilker Erkut, administrative operations assistant at the University of Maryland, quickly learned that after he first applied AI to his work. He had a five-hour deadline to summarize the themes of a book that an executive needed for a discussion. Instead of skimming the texts as fast as he could, he turned to ChatGPT and finished in two-and-a-half hours.
Now, Erkut uses AI every day to learn complex Excel formulas to build spreadsheets that can take polls and track and analyze customer service calls. He uses it for brainstorming, editing and to create custom graphics, something he said he didn’t have the confidence to do on his own. He saves about 15 hours a week, which has allowed him to volunteer for other projects.
“My creativity is higher because I’m focusing on the areas I want to use my own brain for,” said the 26-year-old.
Kanika Khurana also finds AI to be a timesaver. The 30-year-old former head of design at ProCreator, a design agency in Mumbai, uses it for research, to help write up meeting minutes from AI-generated transcripts and to summarize her dictated thoughts after job-candidate interviews, then formatting them for human resources. Her team has also created custom bots that can help other, especially more junior colleagues, create content for clients using specific lingo and voice, and catering to target audiences.
“I’m stretched thin on time, and I often have to take multiple [candidate] interviews,” she said. “After an interview I’d just say what I liked or didn’t like … and say, ‘can you structure it in a proper tone to give to HR?’”
Structuring feedback can be especially helpful when there’s multiple items, discovered Jake Samuelson, co-founder at AI-powered learning platform Uplimit, which is based in San Mateo, California. Samuelson, 40, collected survey data from more than 1,000 people who’ve used his start-up’s platform. He tapped AI to analyze comments, summarize themes and identify areas for improvement. The exercise inspired an AI-powered product feature for instructors to gather similar insights from their students via embedded surveys.
AI also helps him prototype and brainstorm, draft courses with instructors and speed up filling out new client procurement forms by automatically retrieving information from previously answered questions. While Samuelson estimates that he saves up to 10 hours a week and produces better work, he’s also been confronted by AI’s weaknesses, which he keeps an eye on.
“It almost seems eager to please … so it’s extrapolating too much from too little,” he said, adding that he sometimes programs the bots to ask more questions if it doesn’t know something. “It’s about building guardrails and using your own intuition.”
Most superusers say they fact-check anything AI spits out. Becca Chambers, former chief communications officer at Fort Lauderdale, Florida-based ControlUp, a company that helps manage employees’ tech issues, said she uses AI search engine Perplexity to verify answers because it provides links to sources. Despite an occasional blip, she taps AI for almost every work task from deciphering shorthand notes to transcribing podcasts, scraping social media for market trends, tweaking the tone of emails, checking grammar and identifying and captioning video clips.
“I’ve figured out a way to mask my neurodivergence,” said Chambers, 40, who has ADHD. “It’s like having an accommodation that’s inexpensive and easy to use.”
She’s even made a BeccaGPT bot, which she trained with a year’s worth of her writing and podcasts, to help brainstorm and edit content for social media using her voice. BeccaGPT inspired her to create bots for company executives so her team can ghost write for them.
Writing is the main purpose for Stan Sukhinin, founder of financial consulting company Sorso in Austin, who uses ChatGPT. That’s because Sukhinin’s native language is Russian, so typing grammatically correct emails previously took up to 20 minutes. AI also helps him with complex data analysis for revenue projections, using more complex statistics.
While he’s found Anthropic’s Claude to be the best with data, he thinks AI generally isn’t good at crunching and categorizing numbers. So he still does a lot of tasks manually.
“If you don’t tell if it’s wrong, it thinks [the output] is okay,” said Sukhinin, 40. “It made so many mistakes, sometimes with easy questions.”
Beyond introducing errors, AI can cause other issues if workers aren’t careful, experts say. They should avoid inputting sensitive information into tools that don’t have proper enterprise-grade security measures. They could end up stealing ideas or information from others. AI might unintentionally introduce bias. And if workers rely on AI too much, they might miss out on learning skills or risk losing skills they have, said Arvind Karunakaran, assistant professor and faculty of the Center for Work, Technology, and Organization at Stanford University.
While most superusers are still uncomfortable with unsupervised AI, some industry leaders believe autonomous AI is the future. At Salesforce’s recent conference, where it unveiled AI sales agents, Nvidia CEO Jensen Huang suggested that “work is going to be done before you think of it.” For Zoom CEO Eric Yuan, that looks like having a “digital twin” that automatically works alongside people.
“I don’t think humans will always be in charge,” Kate Jensen, Anthropic’s head of revenue, said just two weeks before the company debuted AI agents. “We have to get used to the idea of AI doing tasks.
As for Ross, AI helped them be more empathetic entering tough conversations and avoid trying to “boil the ocean” for every task. It’s even elevated their responsibilities, they said.
“I’ve always been hard to put in a role,” they said. “But now, I’ve gone from supporting everyone to … helping find where the company is going.”
Total Posts: 10261