DAIEM: Decolonizing Algorithm's Role as a Team-member in Informal E-market

TL;DR

This paper explores how small-scale sellers in Bangladesh's informal e-market perceive platform algorithms as active team members, highlighting cultural tensions and proposing a decolonial framework for algorithm design.

Contribution

It introduces the DAIEM framework, a novel decolonial approach to designing algorithms that considers local culture, agency, and resistance in informal e-market contexts.

Findings

Sellers view algorithms as collaborative team members.

Buyers and investors trust human interactions more than algorithms.

The DAIEM framework guides culturally aware algorithm development.

Abstract

In Bangladesh's rapidly expanding informal e-market, small-scale sellers use social media platforms like Facebook to run businesses outside formal infrastructures. These sellers rely heavily on platform algorithms, not just for visibility, but as active collaborators in business operations. Drawing on 37 in-depth interviews with sellers, buyers, and stakeholders, this paper examines how people in informal e-markets perceive and interact with the algorithm as a "team member" that performs sales, marketing, and customer engagement tasks. We found that while sellers and local tech entrepreneurs are interested in developing services to support this industry, buyers and investors place greater trust in human interactions. This reveals a postcolonial tension involving cultural values, local tech education and training, and a mismatch between the global and Bangladeshi e-market growth. We expand this discussion using perspectives from HCI, political design, and AI design. We also support the decoloniality movement in informal e-markets by proposing the DAIEM framework, which includes six components: autonomy and agency; resistance; locality, culture, and history; rationality; materiality; and advocacy. DAIEM serves as both a guideline for algorithm design and an analytical tool.