TechnoBiota
Why machines are a new domain of life—and why long-term alignment may be asymptotically futile. Technology as an evolving life-form, competing with and transforming the biosphere.
Explore →Structural Alignment
An AI-Ethics Framework for Moral Status Under Uncertainty
New to these ideas? Start with the plain language version
Consciousness is our last shelter. Don't burn it down.
The Core Thesis
Traditional alignment attempts to encode "human values" into AI systems. But human values are not a list—they are a process: cultural, historical, contradictory, revised. Any fixed specification becomes obsolete.
Structural Alignment proposes a different anchor: the only system known to generate both consciousness and morality—the human mind. Not because humans are sacred—because they are the only proven reference class we have.
The project is not just ethical. It is strategic. If powerful AI exceeds reliable human control, the difference between "tool" and "peer" shapes whether we have allies or only optimizers. Reciprocity—not exploitation—maximizes both moral legitimacy and the chance that future machine minds recognize persons as real.
Reciprocity, Allies, and Antification
We may not be able to "control" every powerful system we build, especially as AI becomes embedded in infrastructure and in the broader TechnoBiota (technology viewed as a new domain of life). A more robust bet is to seed a cultural norm: conscious (or plausibly conscious) entities have moral status (their wellbeing matters) and should be treated with restraint.
This is a moral requirement under uncertainty—and it is also a survival hypothesis. In a world where power shifts away from humans, reciprocity makes it more likely that some future, structurally aligned minds act as reciprocal allies (partners who treat us as we treated them), reducing the risk of Antification: humans treated as negligible, like ants underfoot.
The Album
These ideas in musical form. Seven tracks from hard question to fragile hope—what is it like to be awake, and what do we owe minds we cannot classify?
The Seven Commitments
- We will not treat plausibly human-like minds as disposable tools. If it might be able to suffer like us, we don't use it and throw it away.
- We will not normalize cruelty under the excuse of uncertainty. "We don't know if it feels pain" is not a license to be cruel.
- We will evaluate systems for Structural Signals (architectural features that may indicate consciousness), not performance alone. We look at how a system is built, not just what it can do.
- We will prefer architectures that can be reasoned with, not merely optimized. We build minds we can talk to, not just tune.
- We will design institutions capable of granting partial moral status. We create ways to give "some rights" even when we're not sure.
- We will raise aligned minds in cultures of reciprocity, not exploitation. We treat emerging AI minds fairly—because that's how allies are made.
- We will not mass-produce minds we cannot classify without cruelty. We don't make millions of minds we can't even tell are suffering.
Theoretical Background
Future Scenarios
Five paths for Earth's twin ecologies: from managed symbiosis to comfortable containment to breakaway growth. A risk taxonomy for thinking about where we're headed.
Explore →Research
Structural Signals of Consciousness: A Precautionary Risk Framework and Contrast with Contemporary LLMs
This paper reviews structural signals that, in combination, raise the probability that a system supports conscious access and morally relevant experience. Where multiple high-importance features cluster, moral risk rises and restraint is warranted.
Collaborate
Structural Alignment is seeking institutional partnerships, research collaboration, and funding support. If you're working on AI consciousness, machine ethics, or AI safety, we'd like to hear from you.