There is a growing belief that scalable and low-cost AI assistance can improve firm
decision-making and economic performance. However, running a business involves
a myriad of open-ended problems, making it hard to generalize from recent studies
showing that generative AI improves performance on well-defined writing tasks.
Research from Rem Koning
Rem Koning
and Rowan Clarke, along with colleagues Nicholas G. Otis, Solène Delecourt, and David Holtz from Haas School of Business at UC Berkeley, delves into the varied effects of Generative AI on the performance of small businesses in Kenya. It explores the differential impact of AI advice on entrepreneurs with varying levels of business success.
The study involved 640 Kenyan small business owners, divided into two groups. One received standard business advice, while the other benefited from AI-generated guidance through a GPT-4 powered mentor. Intriguingly, businesses that were already performing well saw considerable improvement with AI assistance. This highlights AI’s potential as a powerful tool for boosting successful strategies. In contrast, businesses that were not performing as well experienced a downturn when using AI advice. This points to a potential disconnect between the AI’s recommendations and the specific needs or capabilities of these less successful businesses.
This research underlines the importance of a nuanced approach to implementing AI in business settings, especially in developing economies. It demonstrates that the impact of AI is not universally positive and depends significantly on the pre-existing condition of the business. The findings emphasize the necessity of tailoring AI solutions to the unique needs and capabilities of each business to ensure equitable and effective outcomes.