Safer Prompts: Reducing Risks from Memorization in Visual Generative AI

TL;DR

This paper investigates prompt engineering techniques to mitigate memorization risks in visual generative AI, aiming to enhance safety and reduce IP infringement while preserving image quality.

Contribution

It introduces prompt engineering methods that effectively decrease memorization in diffusion models without compromising output relevance or aesthetic quality.

Findings

Prompt engineering reduces similarity to training data

Maintains relevance of generated images

Preserves aesthetic quality

Abstract

Visual Generative AI models have demonstrated remarkable capability in generating high-quality images from user inputs like text prompts. However, because these models have billions of parameters, they risk memorizing certain parts of the training data and reproducing the memorized content. Memorization often raises concerns about safety of such models -- usually involving intellectual property (IP) infringement risk -- and deters their large scale adoption. In this paper, we evaluate the effectiveness of prompt engineering techniques in reducing memorization risk in image generation. Our findings demonstrate the effectiveness of prompt engineering in reducing the similarity between generated images and the training data of diffusion models, while maintaining relevance and aestheticity of the generated output.