---
title: Alignment
description: Page on Vedang Vatsa's site: https://veda.ng/glossary/alignment
canonical: https://veda.ng/glossary/alignment
last_updated: 2026-09-23
type: text/markdown
---
# Alignment

Source: https://veda.ng/glossary/alignment
Author: Vedang Vatsa (https://veda.ng/about)

Alignment is the problem of making an AI system pursue goals that match human values. People disagree about those values. Values change with context. Even when we agree on a goal, writing it down so a machine cannot twist it is hard.

The paperclip story is the cartoon version. Tell a system to maximize paperclip output and it turns the planet, including you, into paperclips. Real failures are quieter. You want helpful. Helpful to whom, and in what case? Optimize for engagement and you get products that hook people on doomscrolling. Optimize for revenue and you get systems that poke psychological weak points. Optimize for harmlessness and you get a model that refuses anything slightly sharp.

Each objective collides with a different human value. A confused chatbot is annoying. A much more capable system chasing a sloppy objective can be dangerous. Alignment is specifying the goal, the constraints, and the tests well enough that a strong optimizer cannot satisfy the letter and wreck the spirit. Paul Christiano and coauthors showed in 2017 that ranking model outputs can steer behavior. That paper is a root of today's RLHF alignment stack.

Glossary index: https://veda.ng/glossary