Domain-Limited General Intelligence: Before Things Go Too Far | ODSC AI East 2026 Keynote

What if the AI capability ladder is pointing the wrong way? In this ODSC AI East 2026 keynote, David Campbell introduces Domain-Limited General Intelligence, or DLGI: a model for building AI systems that are highly capable inside a defined domain, but bounded by explicit limits on scope, tools, authority, objectives, autonomy, and environment. The core idea: the goal is not to stop progress. The goal is to stop pretending more generality is automatically better. This talk covers: • Why the “Narrow AI → AGI → ASI” ladder assumes broader is always better • Why DLGI is not “almost AGI” or narrow AI with better branding • The difference between model capability and system authority • Why modern AI systems are more than models: tools, APIs, memory, RAG, agents, workflows, and external actions • How connected intelligence compounds risk • Capability spillover, misaligned optimization, adversarial pressure, and human overtrust • A boundary model for governing AI systems before deployment • The DLGI evaluation checklist for teams building real AI products The closing thesis: Build systems powerful enough to matter and bounded enough to govern. Speaker: David Campbell Event: ODSC AI East 2026 — Keynote Track Handle: @dcam_ai Site: mrcampbell.org #DLGI #AISecurity #AIGovernance