Concern 204 · Working draft · editorial review in progress
Fine-tune opacity
Fine-tune opacity occurs when an application does not explain which model it builds on or how customization may have changed its behavior.
Understanding the concern
A product may combine a base model with extra training, prompts, retrieval, and software tools. Tests of the original model may not describe that final combination. Customers can mistakenly attribute the base model's documented strengths or safeguards to a modified application.
Using a base model is normal and does not by itself make a product misleading. The concern is whether buyers receive enough information about the actual system to evaluate it. Fine-tuning, adding reference documents, and changing instructions are different modifications and should not be blurred together.
An example
A vehicle is sold using the safety results of its original design after important parts have been changed.
Illustrative unless identified as a documented case.
Words in plain English
- Model
- The learned part of an AI system. It uses patterns found during training to produce an answer or prediction.
- Training
- The process of using examples to shape how an AI model behaves. Learning patterns from material does not establish that the material is true or fair.
- Prompt
- The instructions or information given to an AI to guide its response.
- Fine-tuning
- Additional training used to adapt a model for a task or setting. It can change behavior and needs its own checks.
Evidence and its limits
Model reporting and fine-tuning research support evaluating the changed system; they do not imply that every customization reduces safety.
Mitchell and colleagues · FAT* 2019 · Research proposal
Model Cards for Model Reporting ↗ (opens in a new tab)Proposes documentation of intended use, evaluation conditions and performance; a card's existence is not proof of quality.
Qi and colleagues · Research study
Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To! ↗ (opens in a new tab)Tests safety changes under both malicious and benign customization; accidental weakening is distinct from intentional poisoning.
Broader frameworks and background
These resources provide context. They do not establish every claim in this concern.
A question to discuss
Which parts of the underlying model's evaluation still apply to this customized product?
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.
Trustworthiness
Trustworthiness means giving people good reasons to rely on you. It grows when promises, actions, and evidence line up. With AI, a friendly voice or a famous brand is not enough: people need to know what was checked and what could still go wrong.
In practice, keep promises small enough to test, record mistakes, and tell people when the tool's limits affect them. If a service promises human review, make sure a person really has time and authority to review it.
Ask together: What would give the people affected a sound reason to trust this use?
Accountability
Accountability means being answerable for decisions and taking responsibility for repair. It requires more than saying that the software made the choice.
Name the person or team responsible, keep useful records, and give them authority to correct problems. Make the route to a complaint clear to the people affected.
Ask together: Who will answer for the outcome and help put things right?
Truthfulness
Truthfulness means making claims that match the evidence and being clear about what we do not know. It includes correcting mistakes rather than protecting a polished story.
Check sources, label guesses, and distinguish a trial result from a proven benefit. If AI helped produce a claim, someone still needs to take responsibility for checking it.
Ask together: What can we honestly say, and what still needs checking?
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.