Building AI that adapts to people, not people to AI. Unexpectech explores intelligent systems that learn from context, extend software capabilities, move between digital and physical environments, and support learning and creativity without removing human agency.
Intelligence exists to support people, not make people adapt to the system.
Whenever possible, users should retain ownership of their software, information, and intelligence layer.
Systems should expose what they did and the context used to make important decisions.
Software should be capable of changing as requirements and people change.
New approaches are built, tested, observed, and revised rather than accepted because they sound promising.
Research ultimately needs to become something that can be built, executed, tested, and used.
Project Semilla
What happens when software can participate in its own development? Project Semilla investigates applications that can interpret requested capabilities, modify their own source code, compile the result, inspect failures, correct them, and integrate successful changes.
The current POC has generated functional modules including user-interface components and has been tested across macOS, Windows, and Linux.
Explore Project Semilla →IAm — Intelligent Agent Manifested
IAm explores a persistent intelligent agent that is not tied to one interface. The same conceptual agent can be manifested through physical robotics, virtual environments, and conversational systems while preserving identity and context across those manifestations.
PiDog Lab provides the physical environment for embodied experiments, alongside work involving LP-MUD environments and Alexa.
Explore IAm + PiDog Lab →rAIn — Reinforced Artificial Intelligence Network
rAIn explores AI built around persistent personal context instead of isolated prompts. It is designed around people on the autism spectrum and their trusted support network, combining profiles, authorized collaborators, semantic memory, event capture, documents, routines, and explainable AI assistance.
The goal is continuity and understanding, not replacement of family members, teachers, doctors, or therapists.
Explore rAIn →The creative AI work connects generation with continued human creation. Current projects include an AI music creation platform for songs and albums, and an Interactive Music Studio where students can play instruments, record tracks, combine music, and use AI-generated lyrics as editable material.
Learning Labs are included in AI only when artificial intelligence changes the method: for example AI-assisted drawing-to-3D workflows, AI-assisted lyric generation inside music software, or person-centered learning support. Conventional gameplay behavior alone is not treated as AI research.
Medical AI should operate over traceable context rather than silently inventing context it does not have. The work focuses on structuring information, searching histories, summarizing consultations, explaining terminology, organizing follow-up, and preserving the relationship between an AI answer and the information that produced it.
Detailed work on healthcare interoperability, human and animal health, and the Functional Medicine Matrix belongs to the Medicine & Health research area.
Explore Medicine & Health →