Have you ever wondered how businesses bounce back after major disruptions like natural disasters? In today’s unpredictable world, where events like floods, earthquakes, and pandemics can strike at any moment, the ability of companies to recover quickly—what we call ‘firm resilience’—is more critical than ever. While we often hear about the growing role of artificial intelligence (AI) in boosting productivity during normal times, its true power in times of crisis has remained less explored. Our recent research sheds light on this crucial area, revealing how AI can be a game changer for businesses facing the unexpected.
The Unseen Strength of AI in Crisis
Over the years, AI has transformed various industries, influencing every aspect of business operations—from optimising supply chains to enhancing customer service. For example, General Motors uses AI to predict equipment malfunctions, and Danone Group leverages machine learning to improve coordination across operations, significantly reducing losses. These examples highlight AI’s efficiency in stable environments. However, what happens if the environment is anything but stable? When a natural disaster hits, operations are disrupted, supply chains are broken, and markets react. How can AI help firms withstand these shocks and recover? The answer may lie in the flexibility and insights provided by AI.
Our recent research published in Information Systems Research investigates AI’s role in helping companies, particularly those involved in goods production, build resilience against natural disaster shocks. We define firm resilience as a company’s ability to successfully confront unforeseen challenges and restore normal operations within an acceptable timeframe. As AI algorithms become more sophisticated, they equip companies with unprecedented capabilities to adapt and overcome such challenges. Imagine pharmaceutical firms using AI to predict ingredient delays caused by a hurricane, or retail businesses rerouting shipments based on real-time weather data to avoid disruptions. These are not just hypothetical scenarios; they are becoming reality, as demonstrated by companies like Biogen, which used predictive algorithms to recover quickly from Hurricane Maria’s impact on its Puerto Rico production.
Measuring AI’s Impact: A Deep Dive
To understand AI’s contribution, we needed a robust way to measure both the impacts of natural disasters and a firm’s AI investment. Unlike previous studies that might only consider a company’s headquarters, we meticulously examined county-level operating sites to precisely gauge the extent and intensity of disaster shocks. For AI investment, we analysed a vast database of online job postings, identifying roles that required AI-relevant skills. This allowed us to quantify a firm’s investment in AI capabilities by looking at the cumulative AI-related skills in its workforce.
Our primary focus was to see if AI truly injects resilience. Using a pooled event study approach, we compared firms with different levels of AI intensity during and after natural disasters. The findings are compelling: firms with higher AI intensity experienced more moderate losses and more positive returns compared to peers with lower AI intensity. Specifically, we found that a firm equipping just 2.4% of its total jobs with AI-related skills could recover approximately the full damage to its corporate valuation within a short period after a severe shock.
How AI Works Its Magic: The Production Function
Beyond simply observing the resilience, we delved deeper to understand how AI achieves this. We explored the underlying mechanisms through the lens of an adapted production function. In simpler terms, we looked at how efficiently firms convert inputs like labour and capital into outputs, especially under uncertain conditions. Our analysis revealed that, when faced with uncertainty, firms with higher AI intensity were able to generate more production outputs for each unit of labour or capital invested. This suggests that AI enhances the elasticity of production, making firms more adaptable and responsive to disruptions. This effect remained consistent even after accounting for various other factors and potential biases.
To ensure the robustness of our findings, we addressed the concern that firms investing in AI might already be inherently more resilient. Using an instrumental variable approach, we combined a firm’s baseline task structure with the growth of AI-relevant tasks among its peers. We also controlled for other general IT investments, such as data analytics, cloud computing, and robotics, to isolate the specific impact of AI. Our results consistently pointed to the same direction: AI plays a significant role in bolstering firm resilience.
A Nuance in AI Adoption: The Underperforming Firms
One of the most interesting findings from our research concerns underperforming firms. We discovered that while underperforming firms could potentially benefit ‘more’ from an additional unit of AI investment—meaning AI could make a greater positive impact on their production outputs—their actual realised productivity was often lower than that of their better-performing counterparts. This gap is mainly due to a lack of complementary investments, such as organisational design changes, staff training, or strategic infrastructure improvements. This finding highlights a critical phenomenon: investing in AI technology is not a ‘one-click solution’. To unlock AI’s full potential, it must be coupled with the right organisational strategies, especially for struggling businesses. Simply purchasing software or hiring a few engineers will never be enough.
Beyond Goods Production: Where AI’s Resilience Shines (and Where It Does Not)
It is important to note that our findings, while significant, do not apply universally across all contexts. Our study primarily focused on firms in goods-producing sectors that use AI as an auxiliary technology. We did not observe similar resilience effects in service industries or among firms whose primary business is inventing AI.
Additionally, while AI demonstrated consistent resilience against natural disaster shocks, we did not find the same effects for uncertainty caused by technological disasters like cybersecurity attacks or industrial accidents. This suggests that the mechanisms through which AI contributes to resilience might be specific to the type of disruption and the industry context. Further research is needed to explore these nuances.
Why This Matters: Insights for Managers and Policymakers
Our research not only validates the technological value of AI, but more importantly, provides actionable insights for corporate management and public policy formulation. Firstly, it spotlights AI’s value in uncertain environments, specifically in the context of natural disasters. Secondly, we provide strong empirical evidence of AI-enabled resilience by using both financial market data and accounting-based measures of firm performance. These findings offer a comprehensive picture of how markets react and how firms perform when AI is in play during a crisis. Our work echoes tentative results from previous literature while offering new perspectives on corporate valuations and management practices.
In an increasingly volatile world, understanding how technology can help businesses navigate uncertainty is paramount. Our study underscores that AI is not just a tool for efficiency; it can be a powerful enabler of resilience. For businesses, this means that strategic AI investments, coupled with thoughtful organisational design, can be a vital safeguard against unforeseen shocks. For policymakers and society, it emphasises the importance of fostering an environment where businesses can leverage advanced technologies to build a more resilient future. As AI continues to evolve, its role in helping us adapt and thrive in the face of adversity will only grow.
About the author(s)
Shen Hongchuan is an assistant professor in the Department of Accounting and Information Management of the Faculty of Business Administration at the University of Macau. He received his PhD from The Chinese University of Hong Kong. Prof Shen’s research focuses on the economics of information systems, including information design on online platforms, online advertising, two-sided markets, and the social impacts of artificial intelligence. His research has been published in top-tier academic journals, including Information Systems Research, Nature Communications, and Proceedings of the National Academy of Sciences of the United States of America.
Text & Photo: Shen Hongchuan
Source: UMagazine Issue 32
Academic Research is a contribution column. The views expressed are solely those of the author(s).
