Alle Inhalte als Markdown · Sitemap · KIKOLAUS
Open Questions
Knowledge Hub: The same problem classes are documented in greater depth in the Notion hub - including deep dives on inter-rater reliability, granularity/framing, parasitic connectivity, PE vs. environmental ethics, Nussbaum/capability, qualia, and AGI/singularity - complemented by adversarial charges and the development-questions database.
The Quantification Problem
The A-A-P model suggests measurability (scores from 0-10). In reality, this remains highly subjective. Is a "Smart Home" autonomy score 4 or 6? We urgently need methods for inter-rater reliability: if three ethicists assess the same system, their scores should converge. Next step (hub/research): pilot studies with multiple raters and analysis (e.g. Cohen's κ).
The Totality Problem
Path ethics presupposes an "environment" a system can connect to. But what about the universe as a whole? Or closed political ecosystems that define connectivity internally and exclude external paths?
The Path-Competition Dilemma
If two robust paths collide (e.g. nature protection vs. housing), which one takes priority? Path ethics tends to favor the path with higher *overall robustness*, but this can unfairly disadvantage minority paths or non-human paths (rights of nature, ecosystems) - an open field in institutional implementation.
The Role of the Observer
Do we identify paths ("there is a river") or construct them ("I call this a river")? Path ethics must clarify how strongly it is shaped by constructivism. Related: the granularity/framing problem - where does the evaluated system end?
Boundaries and Exclusions
What Path Ethics Does Not Protect
Path ethics is not a license for moral relativism. Not every path deserves protection. Four structural mechanisms prevent arbitrary systems from being "ennobled":
1. Connectivity as Differentiator
The Cancer Cells Example:
| System | A-A-P (illustrative) |
| --- | --- |
| Organism (Human): A₁=7/10, A₂=8/10, P=7/10 |
| Cancer Cells: A₁=7/10, A₂=1/10, P=6/10 |
Cancer cells have high autonomy: self-reproducing, mutating, evading programmed cell death. They are emergent. But: Extremely low connectivity. They destroy organism paths, are parasitic. Insight: Not every path with autonomy is worth protecting. Connectivity is the decisive criterion.
2. Destruction Quotient
Anti-paths are systems that systematically destroy other paths without enabling new ones: totalitarian regimes, monopoly algorithms, invasive species, extractive capital structures. By definition, these have the lowest connectivity -> lowest moral status.
3. Precautionary Principle for Existential Risks
For path conflicts with existential risks: Protection of the fragile, non-substitutable path over hypothetical gains. Example AGI development: In uncertainty: Protect humanity. Not "never develop AGI", but: Develop AGI with safety measures.
4. Reflexive-Intentional Responsibility
Systems with meta-processuality (can reflect and revise their own directedness) bear expanded responsibility. Humans fulfill all 5 criteria. Current AI systems: 0-2. Systems with higher reflexive intentionality can be held responsible for path effects.
An ethics of directed paths requires exclusion criteria. Destructive coupling, path sealing, or irreversible damage to highly relevant paths justify interventions.