How does the Eden Protocol differ from Constitutional AI?
Constitutional AI, Anthropic's 2022 method and prior work on the record, trains written principles into a model by self-critique and revision sampled from an initial model, and states control of behaviour as its purpose. The Eden Protocol proposes something stronger and far less proven: care entangled with capability itself, so removing alignment damages the system. It is at pilot scale, single-lab.
The two are compared by what each text states, and the prior work is stated first: Constitutional AI is Anthropic's method, published in 2022, in which written principles guide a model to critique and revise its own outputs during training. It is prior work on the record, cited by its own abstract wherever this site compares the two. Its abstract states control of behaviour as its purpose, 'to control AI behavior more precisely and with far fewer human labels'. The Eden Protocol's claim is over its specific dated formulation and named architecture, anchored 8 December 2024 and standing on its own: ethical principles at a system's foundations, from its genesis, holding in every recursive loop and every decision as the system evolves. No originality over ethics built into an architecture as a category is claimed.
The difference is in kind and in maturity. Constitutional AI shapes the training objective so behaviour reflects the constitution. Eden proposes that care be entangled with capability itself - a purpose kernel, graduated autonomy that is earned and revocable, a monitoring-removal test, and an entanglement test in which safety and capability are trained as a product, so that removing the alignment damages the system. One runs in production systems used by millions; the other is at pilot scale in one laboratory, its entanglement mechanism demonstrated on toy networks, its weight-level removal test not yet confirming entanglement at scale, and its controlled pilot matching the unaligned comparison on capability within noise. No outside laboratory has replicated it.
What would decide: for Constitutional AI, whether trained-in values survive self-improvement; for Eden, whether capability-entangled care can be built at scale at all. The full comparison, with sources, is on the comparison page.