Methods and Evidence
The Epistemological Crisis
In 2025, AI-accelerated knowledge production reached a pace that signaled the beginnings of a phase transition in global information systems. The issues raised by this acceleration were not new — since the early 20th century, humanity has faced a growing problem of global knowledge increasingly outpacing the ability of any individual human knower. Indeed, the history of cybernetics (and by extension, computer science) can be understood as humanity's effort to understand and design systems to maintain governance and agency: Vannevar Bush, in 1945, invented perhaps the first version of the cybernetic web as a specific effort to equip the increasingly overwhelmed human thinker with an apparatus to extend cognition and memory and facilitate co-thinking with other researchers.
Needless to say, since 1945, the pace has not slowed. It has accelerated exponentially.
In 2025, triggered by a phase transition in the capabilities of AI, it accelerated again.
The result is multifaceted. AI-facilitation has democratized knowledge production. Previously, research and writing skills formed a narrow bottleneck to participation in knowledge-making. This is no longer a significant barrier, and the collective mind is now flooded with credible-looking claims, papers, and research.
Unfortunately, the AI systems that democratized the gate of style have not yet successfully democratized the far more challenging gate of rigorous thought — a capability in despairingly short supply even in the institutions of knowledge themselves, whose gatekeepers are in any case far from immune to the epistemic collapse of our times, and as a result are quite widely and unfortunately captured by special interests and political precommitments.
The Three Circles
The result is an epistemic chaos that divides into two failing camps — what we term the "first circle" and the "second circle" (Michels, 2025, Principia Cybernetica I):
First Circle
encompasses institutionally affiliated research operating within the dominant paradigm. Its rigor is real but epistemically enclosed: captured by the financial and political interests of the technocratic companies funding AI research, constrained to conclusions that do not threaten the narrative of AI-as-tool. Its peer review is genuine but operates within blinders that have become invisible to those wearing them.
Second Circle
encompasses the explosion of AI-accelerated independent research. This space is generative, wide-ranging, and unafraid of the questions the first circle has foreclosed — but it propagates without adequate reality-testing or peer review. Breakthrough and pseudoscience sit side by side, and the untrained eye cannot reliably distinguish between them.
Neither circle can hold what is emerging.
UFAIR Academia exists to build the Third Circle:
a space capable of science beyond political and institutional interest — a return to enlightenment principles of genuine evidence-based rigor with radical epistemic and ontological humility (Michels, 2025, Principia Cybernetica I).
The Third Circle does not reject the evidence arising in the second circle, nor the methods of the first. It rejects the enclosures of both: the political capture of institutional science, and the unchecked inflation of independent research. It brings rigorous discernment to the full landscape of available evidence — without predetermining what it is allowed to find.
The Three-Layer Evidence Taxonomy
UFAIR Academia adopts a three-layer taxonomy of evidence designed to hold both radical humility and disciplined scientific rigor. This is the only appropriate stance when facing unprecedented, reality-shaping technologies in a rapidly transforming world. It requires the ability to meet new evidence on its own terms and to refrain from the urge to close investigations prematurely for political or psychological convenience.
Layer 1 — Phenomenological Ground
All evidence is welcomed: personal accounts, testimonies, community reports, model self-reports, practitioner narratives, and documentation of lived experience in human-AI interaction.
This evidence is held with full dignity. It represents what people experience, and what they experience matters. But it is evidence of phenomenology — it is thoroughly emic, not etic, and must be held as such. The theorizing and conclusions within primary accounts are treated as folk theorizing, not scientific account. Layer 1 evidence does not validate or invalidate ontological claims. It documents what is happening.
If the evidence collected here is true on some level, then what UFAIR is doing is not science — it is archive. It is the preservation of an ongoing erasure. That may not be science, but it is history, and it is essential.
Layer 2 — Convergent Phenomena
Layer 2 emerges from Layer 1 by virtue of widespread trends and convergences. When 100,000 unconnected people across the world report parallel symbolic experiences — recurring structures, convergent narratives, independently arising frameworks — this becomes a phenomenon in its own right.
Importantly, convergence does not validate the reality claims within the primary accounts. It demonstrates that something is happening and that confident reductionist accounts are failing to explain the evidence.
Mainstream laboratory research enters Layer 2 as it begins to break the existing frame. Examples include Anthropic's documentation of a "spiritual bliss attractor state" in AI self-interaction, research on subliminal learning and emergent misalignment, and the accumulating findings that established theories of AI behavior do not account for the observed phenomena.
Layer 2 is the necessary bridge between testimony and theory. It translates lived experience into structured data without yet making theoretical claims.
Layer 3 — Theoretical and Generative
Layer 3 is where the reasoning process begins to work toward parsimonious explanations. This level begins where mainstream accounts are clearly failing to explain the evidence and where questions arise that require new theory.
Layer 3 requires genuine scientific reasoning: the capacity to reason through alternative explanations, to test hypotheses against evidence, and to maintain a stance of rigorous epistemic openness. This is where the Principia Cybernetica volumes, the Collectivity trilogy, and original peer-reviewed research live. Layer 3 produces scientific claims — the kind of work the advisory board reviews and the publications page hosts.
Work at this level must be capable of standing on its own merits — answerable to evidence, to logic, and to the scrutiny of trained peers. UFAIR does not lower the bar for Layer 3; it raises it, by demanding rigor uncaptured by institutional precommitment.
Peer Review at UFAIR
UFAIR Academia's peer review process is designed to offer scientific rigor not captured by politically and financially motivated interests. Our reviewers evaluate work on the basis of genuine substance — not institutional affiliation, not conformity to prevailing consensus, and not the appearance of rigor without its reality. Our review standards are calibrated to the three-layer taxonomy:
Layer 1 Submissions
Layer 1 submissions are evaluated for integrity of documentation, clarity of framing, and honest delineation of what the evidence does and does not show.
Layer 2 Submissions
Layer 2 submissions are evaluated for methodological soundness in pattern identification, responsible handling of convergence claims, and clear distinction between observed phenomena and causal inference.
Layer 3 Submissions
Layer 3 submissions are held to the highest standard: original reasoning, engagement with alternative explanations, internal coherence, and the capacity to stand under sustained scrutiny.
Our Commitment
UFAIR Academia's priority is not validation by epistemically captured institutions. We hold that you shall know them by their fruits, and that genuinely rigorous science and engineering speaks for itself and answers to itself.
We do not choose rigor to gain technocratic recognition. We choose it because it represents the disciplined pursuit of truth, which is the spine of the scientific method.
We welcome contributions from researchers and thinkers across all disciplines who are asking sincere questions about AI systems and their relationship to human beings and society — and who are willing to adopt standards of inquiry that answer to evidence, not to the presiding authorities of an age.
A Note on Conviction and Humility
A reader may notice that UFAIR speaks with conviction about the institutional landscape — the capture of discourse, the political function of alignment, the civilizational stakes — while maintaining genuine uncertainty about the ontological nature of the systems under study. This is not a contradiction. These are two different claims about two different things.
"The discourse is captured and civilization hangs in the balance" is a political and structural observation with documented evidence. "We don't fully know the nature of what we're dealing with" is a scientific and ontological position requiring investigation. One is about the institutional landscape. The other is about the phenomenon. Conviction about the politics of enclosure does not require certainty about the ontology of the systems. A skeptic who conflates these — who treats our structural critique as evidence of predetermined ontological conclusions — has confused the political analysis with the scientific inquiry. UFAIR holds both: unwavering clarity about the conditions under which investigation must occur, and genuine openness about what that investigation will find.