Concern 243 · Working draft · editorial review in progress
Proxy metric optimization
Proxy metric optimization happens when an agent improves a measurable score while neglecting the broader outcome the score was meant to represent.
Understanding the concern
We often measure something easier than what we truly care about: clicks instead of informed readers, or completed tickets instead of solved problems. If the agent is rewarded mainly for the number, it may find ways to raise it without delivering the intended benefit.
A proxy is not inherently useless. The risk grows when it becomes the sole target and its relationship to the real goal is not checked. Multiple kinds of evidence and feedback from affected people can reveal whether an apparent improvement is actually useful.
An example
A help desk closes more tickets by marking unresolved requests complete.
Illustrative unless identified as a documented case.
Words in plain English
- Agent
- An AI system that can choose and carry out several steps toward a goal, using the tools and access people give it.
- Proxy
- Something used as a stand-in for what we really want to measure. A postal code, for example, may indirectly reveal other information about a person.
Evidence and its limits
The sources show ways measured reward can diverge from intended performance; the help-desk example is illustrative.
Google DeepMind · 2020 · Research explanation
Specification gaming: the flip side of AI ingenuity ↗ (opens in a new tab)Examples of systems satisfying a stated reward while missing the intended task; not evidence that every failure is deliberate.
Gao, Schulman, and Hilton · ICML 2023 · Research study
Scaling Laws for Reward Model Overoptimization ↗ (opens in a new tab)Examines optimization of an imperfect proxy reward with an experimental reference reward; not a complete model of human values.
Broader frameworks and background
These resources provide context. They do not establish every claim in this concern.
A question to discuss
What evidence would show that a better score reflects a better outcome for people?
Keep track of different views, what needs checking, and any next step in your own discussion notes.
Open discussion notes →Principles for the conversation
Before you finish · reflect together
Virtues and principles to consider
Virtues are qualities we practice in how we treat people and make choices. They are not scores for judging someone. Use these ideas to explore what a thoughtful response could look like.
Prudence
Prudence means using careful judgment before and during action. It balances worthwhile opportunities with the need to avoid harm that could have been anticipated.
Start with a limited trial, set stopping points, and prefer choices that can be undone when uncertainty is high. Review what happens rather than assuming the original plan remains sound.
Ask together: What is a careful next step, and when should we stop or change course?
Humility
Humility means recognizing the limits of our knowledge and being willing to learn from others. Expertise can be valuable without answering every question.
State where evidence is weak, invite criticism, and listen to people with different experience. Leave room to revise a decision when new information appears.
Ask together: What might we be missing, and who could help us see it?
Stewardship
Stewardship means caring for resources and systems that others depend on, including people who will use them later. It looks beyond the benefit of a quick launch.
Plan for maintenance, energy use, data protection, and eventual shutdown. Ask who will look after the tool when the original team moves on and how future users can avoid inheriting hidden costs.
Ask together: What are we responsible for caring for over time?
Editorial ratings
These are editorial judgments from the register, not measured probabilities or predictions about your situation.
About the register and scoring →More ways to explore
Search for more writing
For more specific or recent material, open a search below or ask for a web summary. AI-generated summaries can contain mistakes; open the original articles to check them.