Concern 69 · Working draft · editorial review in progress
Multimodal cross-leakage
Multimodal cross-leakage is exposure of learned material across different forms of input and output, such as text prompting an image resembling training data.
Understanding the concern
A model that connects words and images may reveal information in a form different from the one used to request it. Research on image generation has demonstrated recovery of training images in tested systems.
The concern should not be overstated as unlimited access across every format. A similar-looking image is not automatically a recovered photograph, and each claimed channel needs evidence. Privacy review must cover more than the text portion of a system.
An example
A text description could lead an image system to reproduce a recognizable training picture rather than create an unrelated new scene.
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.
- Multimodal
- Able to work with more than one kind of information, such as text, pictures, or sound.
Evidence and its limits
The diffusion-model study demonstrates training-image extraction; it does not prove every proposed cross-modal leakage pathway.
Carlini and colleagues · USENIX Security 2023 · Research study
Extracting Training Data from Diffusion Models ↗ (opens in a new tab)Demonstrates training-image recovery in studied generative models; does not establish every multimodal leakage claim.
Broader frameworks and background
These resources provide context. They do not establish every claim in this concern.
A question to discuss
Were privacy tests run across the system's image, text, and other supported inputs and outputs?
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.
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?
Justice
Justice means fair treatment, fair procedures, and attention to who receives benefits and burdens. It asks whether people with less power can be heard and seek repair.
Look at outcomes across different groups and individual cases. Provide a usable way to challenge decisions and correct harm. Equal wording in a policy does not guarantee equal treatment in practice.
Ask together: Who benefits, who bears the cost, and who can challenge the decision?
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.