Position: General Alignment Has Hit a Ceiling; Edge Alignment Must Be Taken Seriously
This addresses alignment challenges in AI for stakeholders in complex systems, but it is incremental as it builds on existing alignment critiques.
The paper argues that General Alignment, which uses a single scalar reward for diverse human values, fails in complex socio-technical systems with conflicting values, leading to structural, normative, and cognitive issues; it proposes Edge Alignment as an alternative approach with seven pillars to preserve multi-dimensional values and support democratic representation.
Large language models are being deployed in complex socio-technical systems, which exposes limits in current alignment practice. We take the position that the dominant paradigm of General Alignment, which compresses diverse human values into a single scalar reward, reaches a structural ceiling in settings with conflicting values, plural stakeholders, and irreducible uncertainty. These failures follow from the mathematics and incentives of scalarization and lead to \textbf{structural} value flattening, \textbf{normative} representation loss, and \textbf{cognitive} uncertainty blindness. We introduce Edge Alignment as a distinct approach in which systems preserve multi dimensional value structure, support plural and democratic representation, and incorporate epistemic mechanisms for interaction and clarification. To make this approach practical, we propose seven interdependent pillars organized into three phases. We identify key challenges in data collection, training objectives, and evaluation, outlining complementary technical and governance directions. Taken together, these measures reframe alignment as a lifecycle problem of dynamic normative governance rather than as a single instance optimization task.