How Far Should AI Go? Exploring the Workforce Limits of Generative AI
Artificial intelligence is like a cool race car, says Marcus Mossberger, future of work strategist at cloud-based software maker Infor. “Everybody wants it, but have you thought about where you’re taking this race car? Have you built roads to drive it on? What are you going to fuel it with?”Companies have been racing to incorporate AI into their workflows, betting that the tech will make good on its promise to make employees more productive and the business more competitive. But despite the enthusiasm, adoption can be stymied by regulation and risk. In other cases, the speed of adoption can be reckless or unnecessary. In a panel discussion during From Day One’s December virtual conference on pioneering approaches to the future of work, Mossberger and other leaders debated the limits of AI in changing the way we work.AI Adoption by IndustryAdoption tends to be high in the tech industry simply because of its proximity to AI and general openness to change, but in others, like healthcare, regulatory constraints slow the process, says Sumana Srikurmam, who leads HR for the network services division at Tech Mahindra, a global tech services firm.“But even within the tech industry,” she said, “no two organizations will be in the same place at the same time, because the cultures differ, the restrictions may differ, and the stage of growth may be different.”Therefore, move with cautious determination, she says. Compliance is an important part of her job, and keeping up with changing regulations is a complicating factor that is currently multiplying the tools needed. Despite its benefits, there will always be risks, like data privacy, biased language-learning models, and misinterpretation. Many companies are required to deploy even more tools that will mitigate those risks.Marcus Mossberger of Infor, Anita Jivani of Avanade and Sumana Srikurmam of Tech Mahindra spoke with journalist Emily McCrary-Ruiz-Esparza during the webinar (photo by From Day One)Yet a rigid deployment plan isn’t necessarily the solution when it comes to innovative applications, says Anita Jivani, global head of innovation at Avanade, a management consultancy that advises on cloud and AI technologies. “In design thinking, we often think about convergent and divergent thinking. When are we in the problem space? When are we in the solution space? This is one where we have to be in the problem and solution spaces at the same time.” Without flexibility, the problem may change before the solution has been decided.Forward-Thinking Applications in HRHuman resources teams are finding their own applications for AI. Most begin with eliminating administrative tasks. “Tasks most of us don’t want to be doing anyway,” said Mossberger. For instance, “you always have to do payroll processing, which includes reconciliation, so you’re looking for errors and exceptions. Why not AI do that for you?”Agentic AI, or artificial intelligence capable of making decisions on its own, also holds a lot of promise for HR. Being able to “outsource” questions about benefits or PTO or company policies to AI-powered chatbots is freeing many practitioners to spend their time on higher-level reasoning tasks.Assisted authorship is another application Mossberg enjoys. He no longer starts a writing task with a blank page, but with a ChatGPT prompt. Others use AI to track and aggregate employee tasks, “so when it comes time to do your performance evaluation, you have a record of all of the great work that you’ve done,” he said.The Limits Practical and Ethical Limits of Artificial IntelligencePanelists agreed that AI will never be a substitute for human empathy or judgement. As AI gets better at generation, we shouldn’t be tempted to allow it to make decisions on our behalf. Nor should we overvalue the accuracy of its results, Jivani explained. “There’s this view that AI is like a god, an all knowing thing, but we need to re-shift and think of it as a super nerdy, really smart neighbor.” That is: fallible.“Anything that is complex and needs human judgment, any ethical decision making, issues, creativity, innovation–these are things that will need human intervention,” said Srikurmam.Mossberger, who considers himself an optimist when it comes to AI, believes that if we use it wisely, we’ll only have more time to interact with each other.And if AI is giving us back time previously eaten up by tedium, “the question is, what are we doing with that time?” said Jivani. “When a meeting ends 15 minutes early, you could take a walk or make a meal, but what you end up doing is refreshing Outlook. Is that what’s going to happen with the extra time that we have? Are we being intentional, or are we just adding more noise to an already very noisy environment?”Emily McCrary-Ruiz-Esparza is an independent journalist and From Day One contributing editor who writes about business and the world of work. Her work has appeared in the Economist, the BBC, The Washington Post, Inc., and Business Insider, among others. She is the recipient of a Virginia Press Association award for business and financial journalism.