This programme covers three topics from the Consciousness & Intelligence map that are usually treated separately: digital minds, mind uploading, and synthetic consciousness. They are grouped because they share a single load-bearing premise, and separating them tends to hide it. If mind depends on the organisation of a system rather than the material it runs on, then all three are possible in principle and the remaining questions are engineering and evidence. If mind depends on the material, all three are impossible in principle, and no amount of engineering changes that.
That premise — usually called computational functionalism — is an assumption, not a finding. It is the honest starting point for this programme and the thing most often skipped. This programme is a peer to the AIHS feasibility study (S-001), the protomolecule curriculum (P-001) and the engineered origins programme (P-004), and it depends throughout on the Frontier Research briefs on consciousness research, machine consciousness and integrated information theory.
Programme Index
A note on the flags
Every claim in this programme carries one of four honesty flags — the same system used across the Institute — and this is a subject where that discipline matters more than most:
Established Well-understood science. Frontier Active research, partly known. Speculative Theoretical, not yet demonstrated. Handwave Evocative, but without real basis.
The premise, stated once
Established There is no scientific consensus on what makes a physical system conscious, and no accepted test for it. That is not a gap this programme can close; it is the condition under which the programme operates.
Frontier Computational functionalism — that consciousness follows from a kind of computation, so carbon and silicon are interchangeable if the computation matches — is the working assumption of most current research on machine minds, including the theory-derived indicator methods now used to assess AI systems. It is a reasonable assumption. It is not established.
Frontier Two named theories in the same category imply the opposite. Integrated information theory holds that consciousness depends on a substrate's intrinsic cause–effect structure, which means a functional duplicate running on conventional hardware would not be conscious — a direct negative on all three parts of this programme. Orch OR, a minority position, implies a stronger negative still. Neither theory is established either; the point is that the programme's premise is contested from inside the science rather than only by outsiders.
Handwave What follows from all this is a discipline rather than a conclusion. Nothing in this programme should be read as claiming that digital minds are possible, or that they are impossible. Both claims currently outrun the evidence, and confident assertions in either direction are the failure mode this Institute exists to avoid.
Part 1 · Minds that run on computers
Digital minds asks what would follow if artificial systems had experience, and how anyone could tell. Established No current system is a credible candidate and no accepted test exists. Frontier The mainstream research response is the theory-derived indicator method: rather than picking a winning theory, extract computational indicators from several and assess systems probabilistically. Frontier Institutions have begun acting ahead of an answer — AI welfare research programmes, external welfare assessments, and a first dedicated conference in late 2025 — which is precaution under uncertainty, not evidence of machine minds. Speculative Questions of legal standing, and of what it means that a digital mind could be copied, paused or run at varying speed, have no adequate framework anywhere.
Part 2 · Copying a mind to another substrate
Mind uploading is the proposal to reconstruct a particular mind by mapping and simulating its brain. Established The mapping half is real and advancing fast: electron-microscopy connectomics was Nature Methods' Method of the Year for 2025, complete synaptic-resolution connectomes exist for C. elegans and the adult fruit fly, and a cubic millimetre of mouse visual cortex has been densely reconstructed — roughly 120,000 neurons and 523 million synapses, from about a petabyte of imaging data, over nine years. Established The scale gap to a human brain remains around six orders of magnitude in volume. Frontier More fundamentally, a connectome is a wiring diagram, not a running model: synaptic weights, neuromodulation, glial contributions and molecular state are not captured by it. Handwave And whether a successful emulation would be you is a question no engineering result can settle.
Part 3 · Building experience deliberately
Synthetic consciousness is the attempt to engineer experience rather than copy it. Established There is no specification to build against, which is the difference between this and every other engineering topic on the map. Frontier Building to satisfy consciousness indicators would optimise for the test rather than the property, and the field knows it. Speculative Whether construction is easier than detection is unresolved, and there is a case that it is not: a built system would face exactly the same evidence problem as a found one. Frontier The ethics arrive before the capability, because deliberately creating an entity that might suffer is a decision made at design time.
How the three connect
The three are one question approached from three directions. Digital minds asks what would count — the evidence problem. Mind uploading asks can an existing mind be moved — the engineering problem, and the only one of the three with a measurable research programme behind it. Synthetic consciousness asks can experience be built from scratch — the design problem. Progress on any one is bounded by the same missing thing: a theory of consciousness precise enough to adjudicate cases outside the human. That is why this programme leans so heavily on consciousness research, and why the 2025 adversarial collaboration that challenged both leading theories matters more to this subject than any result in machine learning. The same honesty discipline that keeps the origins programme grounded is applied here to a subject with even less settled ground beneath it.