The research that no one
is funding
maybe matters most.
Most knowledge-discovery infrastructure – including the most powerful AI – can only find what has already been asked. That leaves an entire layer of the scientific landscape permanently dark: questions that exist, but have not yet accumulated enough signal to become visible.
GapDetection.com is a systematic method for mapping those questions before consensus forms – in the domains where the absence of research carries the highest human cost.
"AI's biggest potential is in the democratisation of science – but only if it can reach the questions that no training corpus has yet learned to ask." – The GapDetection Premise
Where consensus ends – and discovery has not yet begun
Knowledge gaps are structural, not accidental
GAP–01 · The structural problem
AI learns from what was asked. Not from what wasn't.
Large language models and search engines are trained on consensus. They are extraordinarily good at synthesising existing knowledge – and structurally blind to its boundaries. They cannot detect their own ignorance.
This creates a reproducible failure mode in science: the rarest conditions, the most stigmatised disorders, the most resource-starved research communities – are precisely those least represented in training data. They remain invisible not because they are unimportant, but because no one has indexed the question yet.
GapDetection addresses this at the structural level: systematic prior-art detection at the pre-consensus layer, where the absence of indexed knowledge is itself the signal.
Life at the edges of what is catalogued
GAP–02 · The Ignorance Graph
Mapping what no one has indexed.
The Ignorance Graph is a structured representation of knowledge voids – not the absence of data, but the absence of questions that have been formally asked. It is the inverse of a citation graph: a map of the dark matter of science.
Layer 01
Query-space archaeology
25 years of search behaviour surfaces patterns no LLM training corpus can reproduce: what people search for when there is no authoritative answer to find. The failed search is the signal.
Layer 02
Pre-consensus entity detection
Before a condition, mechanism, or intervention acquires a canonical name, it exists only as a cluster of co-occurring queries and symptom descriptions. GapDetection identifies these clusters before they stabilise.
Layer 03
Systemic gap formalisation
Gaps are not random. They cluster around structural features: funding aversion, diagnostic rarity, geographic inaccessibility, stigma. Mapping the pattern reveals the next gap before it manifests.
Layer 04
Defensive publication
Every identified gap is documented as open prior art – strengthening the public prior-art record – reducing the risk of exclusive enclosure and keeping the knowledge infrastructure open.
"The most important scientific questions are often those that no one has yet thought to fund – because the community that would benefit from asking them lacks the institutional weight to make them visible."
– GapDetection.com · On the ethics of structured ignorance mapping
GAP–02b · The Ignorance Graph
The inverse of everything AI can currently see.
A citation graph maps what is known and how knowledge connects. The Ignorance Graph maps the structural absences – the entities, questions, and causal chains that exist in reality, but have not yet accumulated enough indexed signal to become computable.
Why this matters for AI
Pre-consensus entities are AI’s blind spot
Before a rare condition has a canonical name — before it appears in MeSH, ICD, or any training corpus — it exists only as a cluster of co-occurring symptom queries and failed searches. Today’s AI cannot reason over it. The Ignorance Graph is the infrastructure to change that.
The detection signal
Failed searches are the primary data
A search that returns no authoritative result is not noise — it is evidence of a gap. Longitudinal search observation surfaces the patterns: what communities search for when no institution has yet accepted the question. The ignorance is the data.
Scale of the invisible
A measurable share of all searches find nothing authoritative
That is not a retrieval problem. That is a knowledge-production problem. The queries exist. The need exists. The research — and its funding — does not. GapDetection maps this layer systematically, domain by domain.
The compounding effect
Gaps not detected, persist
Without formal detection and documentation, a research gap remains invisible to funding bodies, developers, and AI training pipelines alike. The Ignorance Graph breaks this cycle by making absence computable — and therefore actionable.
“The most consequential scientific questions of the next decade are not in the literature yet. They are the ones not yet formally asked — because the communities who need them answered have never had the institutional weight to ask them at scale.”
— ignorancegraph.com · On structural knowledge voids
GAP–03 · Five open domains
The infrastructure spans five open domains – plus grey-mass areas.
Each domain represents a research space where the gap-to-consequence ratio is highest: where the absence of indexed knowledge directly translates into delayed diagnosis, missed treatment, or preventable suffering.
Health
Gaps in established health domains – conditions with high prevalence but under-researched mechanisms.
02 / 07Therapy
Mental health and behavioural intervention spaces where apparent solution saturation masks unresolved sub-conditions.
03 / 07Rare Diseases
7,000+ conditions. Fewer than 5% with approved treatment. The highest-density gap landscape in biomedical science.
04 / 07Science
Cross-disciplinary detection of knowledge gaps at the intersection of established fields – where neither community has claimed the question.
05 / 07Research Fundraising
Structured mapping of the gap between research need and funding landscape – identifying where philanthropic and institutional capital is missing.
GAP–04 · Methodology
Systematic. Reproducible. Defensively open.
-
Search behaviour analysis
Longitudinal search observation over 25 years reveals what people seek when no authoritative answer exists. The failed search is the primary signal – not what was found, but what was not.
-
Systemic gap formalisation
Identified voids are formalised using systems-thinking frameworks – mapping the structural conditions (funding, access, stigma, rarity) that produced the gap and that sustain it.
-
Defensive prior-art publication
Every detected gap is documented in the public record, strengthening the prior-art foundation and reducing the risk of exclusive enclosure. The knowledge infrastructure is open by design – not by accident.
-
Research community activation
Formalised gaps are shared with the relevant communities: patient groups, academic researchers, underfunded labs in low-resource settings – those who need the signal most.
-
Knowledge Graph integration
Structured gap data is prepared for integration with entity-based knowledge systems – linking pre-consensus entities to existing taxonomies and enabling AI-scale reasoning over the ignorance layer.
-
The Ignorance Graph
The culminating layer: a persistent, cross-domain map of structural knowledge voids – all detected gaps formalised into a queryable graph that makes the invisible computable. Where the Knowledge Graph ends, the Ignorance Graph begins.
GAP–05 · The unique edge
Three disciplines. One convergence.
Gap detection at this level of precision requires a specific combination of capabilities that does not arise from a single discipline – or from a single career. It requires thirty years of cross-domain signal accumulation in direct contact with the problem space.
Structural observation
Illuminating the invisible
Dimension 01 · 30+ years
Marketing intelligence
Three decades of attention-architecture analysis – understanding how communities form around problems, how language evolves before consensus, and how demand surfaces before supply is aware of it.
Dimension 02 · 18 years
Systems thinking
Formal training in organisational systems thinking applied to research ecosystems – mapping the feedback loops, structural delays, and resource dynamics that explain why certain gaps persist across decades.
Dimension 03 · 25 years
Lived brain research – from medical necessity
Epidermoid tumour – a rare, slow-growing, often misdiagnosed intracranial condition with fewer than 1,400 documented English-language research entries before 2010. Twenty-five years of navigating a fragmented, under-resourced research landscape from the inside. The knowledge gaps in this domain are not theoretical. They are the origin of this methodology.
GAP–06 · Ethical infrastructure
Research that cannot be locked away.
Every design decision in this infrastructure is made against a single criterion: does it increase or decrease the probability that a researcher in a low-resource setting, working on a neglected condition, can access and build on this work?
Open prior art
All identified gaps are published as defensively open prior art – strengthening the public prior-art record, reducing the risk of exclusive enclosure, and ensuring they remain a shared scientific asset.
No budget dependency
The detection methodology requires no grant approval, no institutional affiliation, and no equipment. It is designed for researchers who have none of these – and need the signal most urgently.
Salutogenic focus
Priority is given to conditions where the research gap most directly causes unnecessary suffering – not where commercial return is highest. Rare diseases, orphan conditions, stigmatised disorders.
Democratised access
Results are shared directly with patient communities, under-resourced labs, and independent researchers – not only with institutions that already have access to proprietary databases.
Research in service of human life
The most fundamental measure of any scientific infrastructure is not citation density or commercial throughput — it is the reduction of preventable suffering. Every research gap that is detected and made visible is a life made slightly more possible. GapDetection operates from the conviction that science is never neutral: the choice of which questions to ask, and for whom, is itself a moral act. Orienting that choice toward the highest unmet human need — toward rare conditions, stigmatised disorders, and populations whose suffering remains poorly indexed — is not a constraint on scientific ambition. It is its highest expression.
GAP–07 · Collaboration
This infrastructure is ready to scale.
The detection methodology, the domain infrastructure, and the open prior-art framework are in place. What scales the impact is integration with AI systems that can reason over pre-consensus entity spaces – and with the scientific communities that need the signal.