Insights from Harvard, MIT, and Wharton Highlight the Risks of Entrusting AI Training to Junior Staff

Insights from Harvard, MIT, and Wharton Highlight the Risks of Entrusting AI Training to Junior Staff

If you’re interested in the latest updates and exclusive content on leading AI trends, join our daily and weekly newsletters.

Many believe that in the race to adopt artificial intelligence (AI), younger, tech-savvy employees will lead the way in teaching their managers how to use these powerful tools effectively. However, a recent study challenges this assumption, especially regarding the fast-changing field of generative AI.

This study, a collaboration between researchers from Harvard Business School, MIT, Wharton, and other institutions, along with Boston Consulting Group, discovered that junior employees who experimented with a generative AI system often made recommendations to mitigate risks that contradicted expert advice. The findings indicate that companies can’t rely solely on reverse mentoring for the responsible use of AI.

The researchers interviewed junior consultants about handling AI-related risks and found that their advice was generally off the mark. Interestingly, the tactics these junior consultants suggested to calm their seniors’ concerns were contrary to the recommendations by GenAI technology experts at the time. This implies that junior professionals might not be the most reliable source of expertise in using this emerging technology effectively for their more senior colleagues.

In mid-2023, the researchers interviewed 78 junior consultants who participated in an experiment using GPT-4, a powerful generative AI system, to solve business problems. Despite lacking technical AI expertise, these consultants shared tactics they would recommend to address managers’ concerns about risks. However, the study revealed that their tactics often stemmed from a superficial understanding of the technology’s capabilities. Their suggestions focused on changing human behavior rather than designing better AI systems and were typically project-level interventions instead of organization-wide solutions.

The researchers pointed out that to understand why junior professionals might not provide reliable expertise on emerging technologies for senior members, one must consider the threat to status and the risks to valued outcomes. Given AI’s rapid development, superhuman capabilities, and heavy data reliance, this is especially relevant.

As companies navigate the opportunities and challenges of generative AI systems, which can engage in open-ended dialogue, answer follow-up questions, and assist with writing, analysis, and coding, this study highlights the limitations of relying on younger employees to guide AI implementation from the bottom up. The study stresses the importance of top-down AI governance, expert input, and upskilling at all levels of an organization.

Senior professionals have the dual responsibility of swiftly adopting new technologies and anticipating future advancements and their implications for both clients and their organizations. To lead effectively in this rapidly evolving technological landscape, seniors need to develop a deep understanding of new technologies and their associated capabilities.

Stay updated with the latest news by subscribing to our daily newsletter.