IGI Research / Research agenda
Better questions.
Testable directions.
A research agenda for alignment, memory, agent security, collective intelligence, consciousness and quantum computing.
An agenda of questions, not a catalogue of breakthroughs.
IGI Research is the research-facing part of Humanity & Intelligence. At this stage, it sets out questions, proposed methods and the evidence we would need before making stronger claims. It does not imply a staffed laboratory, active studies, institutional partnerships or completed experiments.
Alignment & control
How can stated goals, actual behaviour and enforceable limits remain consistent?
Compare behaviour across ambiguous tasks, incentives and oversight conditions. Test whether a system exposes conflicts or works around constraints.
Memory & continuity
Which information should an intelligent system retain, and who controls it?
Compare stateless and governed-memory workflows. Inject stale, contradictory and untrusted records; inspect correction, provenance and deletion behaviour.
Multi-agent systems & security
Does delegation improve the task without making authority harder to understand?
Use explicit task contracts, separate tool permissions and a record of delegation. Test shared errors, misleading messages and recovery after a component fails.
Collective intelligence & human oversight
When do people and AI perform better together?
Compare with the strongest relevant baseline. Measure error detection, calibration, review effort and outcomes for affected people.
Consciousness & emergence
What would count as evidence, and which alternative explanations remain?
Separate behavioural performance, self-reports and theory-based indicators. State what an experiment can discriminate before interpreting its result.
Quantum computing & intelligence
For which well-specified tasks could quantum methods provide a useful advantage?
Compare a defined quantum or hybrid approach with strong classical methods, including data access, noise and resource requirements.
Two distinctions that deserve care.
Butlin and colleagues’ 2023 report derives possible indicators from scientific theories of consciousness. This is a structured research approach rather than a universally decisive test. The question of experience should be kept separate from how much authority a system may exercise.
IBM’s quantum-computing overview describes quantum machine learning as an active area with open questions about practical advantage. In this agenda, quantum computing is a topic for careful comparison, not an explanation of IGI or evidence of a quantum product.
How a Deep Dive earns its place.
A precise question.
Define terms and describe the decision the analysis could change.
Primary sources and strong objections.
Show what was actually studied, the comparison used and the limits of the result.
A visible evidence status.
Label findings, competing interpretations, hypotheses and IGIQANT proposals separately.
A test or a reason to remain uncertain.
Explain what observation could weaken the argument. Preserve uncertainty where evidence cannot decide.
Future Deep Dives will be added as complete, reviewed pieces. The topics above are an agenda, not links to articles that do not yet exist.
Read the first analysisSources & editorial method
- AI ControlGreenblatt et al.
Research on safety protocols under intentional subversion in a specified programming-task setting.
- AI Risk Management FrameworkNIST · AI RMF Core
Voluntary risk-management framework; no claim of IGIQANT certification.
- Excessive agencyOWASP · LLM06:2025
Security guidance on functionality, permissions and autonomy.
- Human–AI collaborationVaccaro, Almaatouq & Malone · 2024 meta-analysis
2024 systematic review and meta-analysis; results depend on the tasks and systems studied.
- Consciousness in AIButlin et al. · 2023
2023 interdisciplinary report proposing theory-based indicators; not a conclusive consciousness test.
- Quantum computingIBM Quantum Learning
Technical educational source describing opportunities and open questions in quantum machine learning.
Prepared with AI assistance for IGIQANT. Original source review: 10 September 2026. Presentation and selected references reviewed on 15 September 2026. Research findings, attributed positions and IGIQANT proposals are distinguished throughout. No endorsement by cited researchers or institutions is implied. Corrections and substantive counterarguments are welcome at contact@igiqant.com.