Social by Nature: How Socio-tecture Shapes the Work of SMBs and Considerations for Reimagining Collaborative Human-AI Systems

Research Background and Issues

• Identified Problems or Challenges:

  • The COVID-19 pandemic has catalyzed the digital transformation of small and medium-sized businesses (SMBs) globally. However, SMBs in Africa, particularly in Kenya, exhibit distinct pathways and challenges in their digitalization process compared to developed countries.
  • Existing research and technological development often adopt a "deficit discourse," portraying Africa as digitally underdeveloped while neglecting the rich socio-cultural context and complex practices of SMBs in the region.
  • Large generative AI technologies are predominantly trained on data and experiences from the Global North, reflecting design assumptions that are misaligned with the context of Kenyan SMBs. This misalignment poses potential risks, including exacerbating the "AI digital divide."

• Significance of the Problem:

  • SMBs account for over 90% of businesses in Africa and provide 80% of employment, playing a critical role in economic development. Understanding the unique needs of African SMBs is essential for global technology design and economic advancement.
  • If technologies like generative AI continue to be developed based on Global North value systems, they may worsen issues of technological applicability and digital inequality.

• Research Motivation and Related Work:

  • The authors aim to reexamine the digitalization process of African SMBs through the "socio-tecture" framework, focusing on social networks, relationship orientation, and the role of human knowledge.
  • The study seeks to challenge the current deficit-based narratives of technology adaptation in Africa and propose recommendations for Afro-Centric technology design.

Proposed Solution

• Proposed Method or Solution:

  • Reframe the digitalization experiences of African SMBs using the "socio-tecture" framework.
  • Analyze how generative AI can serve local businesses by integrating the scenarios and values of Kenyan SMBs, rather than simply importing tools developed in the West.
  • Propose innovative concepts for generative AI design, emphasizing three principles: "dependence on social networks," "relationships over transactions," and "humans as carriers of knowledge."

• Innovative Aspects of the Solution:

  • Redefine technology adaptation and design issues through the specific socio-cultural lens of socio-tecture, rather than relying on traditional design logic rooted in "WEIRD" (Western, Educated, Industrialized, Rich, Democratic) contexts.
  • Challenge the current user-centric design philosophy of generative AI, which focuses on individuals or independent users, by emphasizing trust in social networks, local expertise, and the value of collaboration. The authors propose designing "community-identity AI."
  • Introduce new concepts such as "relationship-oriented AI" and "collaborative generative AI," highlighting that AI design should support business relationships, flexibility, and equity rather than solely pursuing efficiency.

• Implementation Steps and Key Technologies:

  • Conduct in-depth observations and background research on seven SMBs located in Nairobi, Kenya (e.g., interviews, field observations) to understand how SMBs use digital tools from a localized perspective.
  • Apply the socio-tecture framework to analyze their daily practices and explore how AI design can support data management, customer relationships, and employee collaboration.
  • Utilize AI-supported research tools (e.g., HeyMarvin) to facilitate rapid data processing and thematic analysis.
  • Propose design principles such as embedding local knowledge, safeguarding privacy and trust, and enabling AI to collaborate dynamically with communities.

Research Outcomes

• Specific Findings:

  • Provided an in-depth description of how Kenyan SMBs approach digitalization based on three socio-tecture principles: "social networks," "relationships over transactions," and "humans as carriers of knowledge."
  • Highlighted how low-cost tools like WhatsApp bridge the gap between the physical and digital worlds. Despite challenges in information management, these tools are widely adopted due to their suitability for relationship-building.
  • Proposed that generative AI design should adopt community-centered strategies to support localized adaptation rather than serving individuals in isolation.

• Comparison with Existing Solutions and Advantages:

  • Unlike traditional AI tools based on Global North datasets and design logic, this approach embeds local traditions and values to propose a more flexible, inclusive, and culturally adaptive design pathway.
  • Suggested "contextualization" and "collaborative adjustment" of AI to align with the practical needs and operational models of African SMBs.
  • Demonstrated the potential of AI to support data management, reduce cognitive load, and enhance business collaboration.

• Experimental or Evaluation Results:

  • The study identified the collaborative burdens SMBs face when integrating digital tools. For example, while WhatsApp is convenient for knowledge management, it lacks formalization and standardization.
  • Collaboration with generative AI could provide short-term benefits by reducing repetitive tasks and enhancing knowledge sharing. However, its long-term value depends on deeply considering local contexts and relationships.

• Limitations and Future Directions:

  • Limitations:
    • Kenya's socio-political environment (e.g., the Worldcoin incident) may have influenced the scope of the study and the authenticity of participant behavior.
    • Observation interference and potential cultural biases of the researchers could also be influencing factors.
  • Suggested Future Directions:
    • Further research on how generative AI can embed and learn from local knowledge ecosystems.
    • Broader cross-cultural experiments to examine the applicability of this approach to SMBs in other African countries and beyond.

Through this consistent analytical approach, the study highlights the critical role of socio-cultural and local values in the application of generative AI technologies. It provides a foundation for building fairer and more inclusive AI systems. This approach holds profound implications not only for Africa but also for the global adaptation of AI applications to diverse cultural and contextual settings.