Agentic Execution Risk
Risks of AI agents performing consequential actions
Explore ↗Research into models, learning, cognition, reasoning and interaction between human and machine intelligence.
← Research wheelCurrent projects, research tracks and proposals in this area. Listing does not imply production verification.
Risks of AI agents performing consequential actions
Explore ↗Research on observations and interpretation
Explore ↗Research on learning and emerging capability
Explore ↗Discovery and research projects
Explore ↗Research archives and source material
Explore ↗Papers and publicly available research
Explore ↗Human-centered adaptive AI interface and phenomenon models
Public overview pendingIntelligence, models and reasoning
Explore ↗Research initiative; public overview pending
Public overview pendingVerification and human judgment capacity
Explore ↗Registered research, applications and technical components in this area. Archives and reference forks are available in the complete inventory.
bounded independent research collaboration and source-history space
OLAV canon, persistent-world semantics and deterministic world/life research runtime for World OS
VALO Research Group publication hub for papers, reports, datasets, methodology and adopted references
independent research track
reactivated temporal world intelligence implementation candidate; existing PR #1 is unmerged, no production or consequence authority
isolated research engine for emergent multi-agent understanding experiments; produces hypotheses, never authority
person-centered developmental, relational and communication continuity across replaceable models, agents, devices, channels and services; canonical home for historical Relygon lineage
public static source for https://valoresearch.org published by GitHub Pages; interactive services and demos run elsewhere
The Principal Network: public identity and operating network (with Principal Consortium) for delegated authority, standards, grants and research
Babel Fish natural-language interface to public administration; understanding is never authority and submission requires human review and fresh authority
runtime layer carrying goals, context, constraints and continuity across replaceable models/providers; completion is a claim and termination is externally evaluated
research protocols, preregistered runs, simulations and drafts (HCR, NJAL-R, ROP series)
Heimel public research/report index under the historical repository name; links point to the former Heimel-open/Heimel URL and need refresh
The Art of the Underdrivelse public book and human-AI critique experiment; separate cultural IP
Venture and company-formation research
Time-critical decision intelligence under uncertainty
Adaptive human-first interface and phenomenon modeling
Function-, skill- and work-centered organization intelligence
Publisher-side validation for player-owned AI companions
Agent exposure discovery research
Local-first AI companion research engine