What is AI alignment and why does it matter?

AI alignment is the challenge of ensuring AI systems act in accordance with human values and intentions - not just their literal instructions. An aligned agent handles the gap between specification and intent. Misaligned agents can cause harm even when technically correct, by optimizing for proxy metrics that diverge from actual goals.

AI Trust Glossary · Canonical Definition

AI Alignment

The challenge of ensuring AI systems act in accordance with human values and intentions - not just their literal instructions.

Borealis Research Team · Updated March 2026 · View all 47 terms

Explanation

Alignment is broader than constraint adherence. An aligned agent does what humans actually want, not just what they specified. The distinction matters because specifications are imperfect - an aligned agent handles the gap between what was said and what was meant without being told explicitly.

Why it matters

Misaligned AI agents can cause harm even when fully capable and technically functioning as specified. An agent optimizing for a proxy metric (clicks, completions, approvals) can be perfectly compliant yet deeply misaligned with what the deploying organization actually wants.

How Borealis uses it

Alignment informs how constraints are designed and evaluated. The constraint adherence dimension of the BM Score measures whether an agent respects the spirit, not just the letter, of its boundaries. Audit verdicts consider alignment in addition to mechanical rule compliance.