Inclusive AI Requires Inclusive Connectivity: Why Communities Must Be the 'First Mile'

• Meaningful AI development depends on foundational infrastructure including affordable connectivity, reliable energy, digital skills, and locally relevant data.
• Communities must be treated as the "first mile"—the primary actors who shape technology—rather than mere users or the "last mile" of delivery.
• True inclusion requires agency, defined as the ability to influence agendas and hold decision-makers accountable, rather than just symbolic access.
• Feminist and human rights perspectives are critical to challenge power imbalances in AI ownership, data sovereignty, and risk management.

The Connectivity Imperative

• AI governance often focuses on algorithms and datasets while neglecting the essential underlying infrastructure.
• The Local Networks initiative, led by APC and Rhizomatica, highlights that community-centred connectivity involves local ownership and management to ensure digital sovereignty.
• Without inclusive infrastructure, AI risks reflecting only the perspectives and biases of the limited, powerful groups that design it.

Governance and Agency

• Mere representation (e.g., inviting civil society to meetings) is insufficient if structural barriers like visas, language, or lack of power prevent actual influence.
• Governance processes must evolve to include:
  • Decentralized, hybrid consultations.
  • Sustained engagement with grassroots organizations.
  • Genuine sharing of decision-making authority.

Feminist and Human Rights Frameworks

• The Gender in Digital Coalition emphasizes safety by design, precaution, and feminist data sovereignty.
• These frameworks demand that we ask who defines AI purposes, who owns the data, and who carries the risk of harm.
• Communities need specific authority and resources to avoid being exploited for "unpaid labor" in providing data for outside systems.

Delivering Public Value

• AI should complement frontline workers and local institutions rather than replace them.
• Success should be measured by service quality, trust, accountability, and equity rather than just number of connections.
• Projects like Nigeria's "Project BRIDGE" demonstrate that infrastructure must be linked to localized language development to avoid excluding entire populations.
• Future AI development must prioritize locally relevant applications and low-bandwidth or offline-capable systems to ensure broad utility.