AIFOD Geneva Summit Panel Explores Decentralised Architectures for AI

The future of artificial intelligence will likely shift from massive, centralised systems toward interconnected networks of specialised tools and local computing, according to panellists at an event organised by the AI for Developing Countries Forum (AIFOD), which publishes this site. Speaking on 14 August at the AIFOD Geneva Summit 2026, international experts during Dialogue XVI considered whether technological progress should be defined by model scale or by real-world reach and community relevance.

Networked Systems and Distribution Challenges

Philip Leaper of SAP argued that emerging ecosystems will rely on diverse knowledge sources rather than single dominant models, noting that the coming landscape will be networked rather than monolithic. Panellists noted that smaller, distributed systems offer practical advantages for developing regions facing infrastructure and connectivity constraints. Dr Nagesh Sharma of ReWiseEd AI warned of artificial inequality, observing that as machine intelligence becomes cheaper, human judgement will become increasingly vital to verify outputs. Sreeraj P.M. of Protean agreed, noting that rapid information retrieval still requires people to evaluate validity.

Integrating Local Knowledge and Adapting Work

Tijn van der Zant of The AI Guys advocated for innovation hubs in schools and universities to help communities integrate local insights, such as agricultural practices absent from global training data. Wolfgang Pinegger of GLMall addressed data ownership, maintaining that responsible AI relies on giving data holders authority over their information so that smaller enterprises remain visible.

Regarding employment, the panel weighed how automation might alter career progression if entry-level work diminishes. Dr Sharma suggested framing labour changes as a choice regarding what human roles should not be delegated. Concluding the session, moderator Rahul Ghatalia stressed that the trajectory of AI will be governed by choices regarding infrastructure, data rights, and inclusion.