As firms race to adopt artificial intelligence systems, it is mostly assumed that Younger, more technically savvy employees will take the lead to show their managers effectively use the powerful latest tools.
But one latest study challenges this assumption in terms of the rapidly evolving technology of generative AI.
The Researchconducted by researchers from Harvard Business School, MIT, Wharton University and other institutions in collaboration with Boston Consulting Groupfound that junior employees experimenting with a generative AI system made risk mitigation recommendations that contradicted expert advice. The results suggest that firms cannot depend on reverse mentoring alone to make sure responsible use of AI.
“Our interviews revealed two findings that contradict the prevailing literature,” the authors write. “First, the tactics that juniors really useful to allay their supervisors' concerns contradicted the recommendations of the experts in GenAI technology on the time. This indicated that juniors might not be the perfect source of experience within the effective use of this latest technology for more experienced members.”
Junior consultants struggle to mitigate AI risks in GPT-4 experiment
The researchers surveyed 78 junior consultants in mid-2023 who had recently participated in an experiment during which they got access to GPT-4, a robust generative AI system, for a business problem-solving task. The consultants, who lacked technical AI expertise, shared tactics they’d recommend to allay managers' concerns in regards to the risks.
However, the study found that junior executives' risk mitigation strategies were often based on “an absence of deep understanding of the capabilities of the brand new technology,” that they focused on changing human behavior slightly than developing AI systems, and that they focused on project-level interventions slightly than company- or industry-wide solutions.
Overcoming the challenges of introducing generative AI into firms
“To explain how and when junior employees might not be a source of experience in using a brand new technology for more experienced staff, we must consider not only the threat to their status but additionally the risks to invaluable outcomes,” the researchers write, pointing to AI's exponential rate of change, its superhuman capabilities and its reliance on vast amounts of knowledge.
The findings come at a time when firms are grappling with the opportunities and challenges of generative AI systems that may engage in open dialogues, answer follow-up questions, and assist with writing, evaluation, and coding tasks. By highlighting the constraints of counting on digitally native employees for bottom-up AI implementation, the study underscores the necessity for top-down AI governance, expert input, and education in any respect levels of the organization.
“Leaders face the twin responsibility of rapidly adopting latest technologies today while concurrently anticipating future versions of technologies and their impact on their customers and their very own organizations,” the authors noted. “To lead their teams and organizations in navigating a rapidly expanding technological frontier, leaders must develop a deep understanding of recent technologies and the opportunities they present.”
