Institutional Costs of AI for Small Retail

Institutional costs surrounding AI may be disproportionally higher for small retail businesses in non–pronoun-drop societies.

Is a small jewelry shop in Naples, Italy, at a disproportionate non-market AI disadvantage simply because it is a small retail business in Italy? As AI tools become increasingly embedded in retail operations, small firms like this often function within incomplete institutions within the broader AI ecosystem. Their technological awareness is underdeveloped compared to the large firms that shape AI policy and adoption. Dorobantu et al. (2017) explain that in incomplete institutional environments, such as AI's fragmented adoption, costs are broadly spread across firms because the “rules of the game” are weak or underdeveloped. Alternatively, Dorobantu et al. (2017) describe captured institutions as settings where political and economic power is concentrated among dominant actors. In this sense, AI is simultaneously incomplete and captured: the rules are still forming, yet large firms such as Apple, Meta, and NVIDIA engage in transformative nonmarket strategies, including lobbying and coalition building, to shape those rules in their favor (Associated Press, 2025). Smaller firms, like the jewelry shop, have no comparable influence and must rely on adaptive nonmarket strategies, adjusting reactively to institutional change (Dorobantu et al., 2017).

Shoham et al. (2024) add a cultural-linguistic perspective, showing that societies whose dominant languages allow first-pronoun drop tend to promote collective responsibility and higher ethical norms. In such environments, costs and responsibilities are distributed more collaboratively. In contrast, in non–pronoun-drop societies such as the United States, the linguistic emphasis on individual agency aligns with more self-interested, competitive behavior (Shoham et al., 2024). This orientation may support faster innovation, as seen in AI-related innovation, but can also amplify inequality.

This pattern connects to my research on AI knowledge and agility in small retail firms. Because resources are limited and adaptive nonmarket strategies impose additional costs, these businesses must treat AI literacy as a core strategy. I hypothesize that the greater a small retailer's knowledge of AI compared to its competitors, the stronger its ability to remain agile and to offset institutional disadvantages.