Ethical-Lens: Curbing Malicious Usages of Open-Source Text-to-Image Models
This addresses ethical concerns for users and developers of open-source text-to-image tools by curbing misuse, though it is incremental as it builds on existing alignment methods without internal model revision.
The paper tackles the problem of malicious misuse of open-source text-to-image models by introducing Ethical-Lens, a framework that refines user commands and rectifies outputs to ensure value alignment, achieving alignment capabilities comparable to or superior to commercial models like DALLE 3.
The burgeoning landscape of text-to-image models, exemplified by innovations such as Midjourney and DALLE 3, has revolutionized content creation across diverse sectors. However, these advancements bring forth critical ethical concerns, particularly with the misuse of open-source models to generate content that violates societal norms. Addressing this, we introduce Ethical-Lens, a framework designed to facilitate the value-aligned usage of text-to-image tools without necessitating internal model revision. Ethical-Lens ensures value alignment in text-to-image models across toxicity and bias dimensions by refining user commands and rectifying model outputs. Systematic evaluation metrics, combining GPT4-V, HEIM, and FairFace scores, assess alignment capability. Our experiments reveal that Ethical-Lens enhances alignment capabilities to levels comparable with or superior to commercial models like DALLE 3, ensuring user-generated content adheres to ethical standards while maintaining image quality. This study indicates the potential of Ethical-Lens to ensure the sustainable development of open-source text-to-image tools and their beneficial integration into society. Our code is available at https://github.com/yuzhu-cai/Ethical-Lens.