Show HN: A concept to make images resilient to AI manipulation

This project introduces an innovative conceptual framework called the Noise Adaptive Algorithm, a method for protecting digital images and videos from unauthorized use and manipulation by AI generative models (such as those used for deepfakes, AI-driven content generation, or training data extraction). The core principle is to achieve this protection without degrading the visual quality for human perception.

With the rapid advancements in generative AI (like Google's Veo and similar technologies), the need to protect personal photos, artistic creations, and sensitive visual data is more urgent than ever. This project aims to contribute a novel approach towards responsible AI usage and data protection.

• Intelligent Noise Embedding (via Adaptive Spider Web): Embeds a specially structured, dynamic noise pattern designed to disrupt AI processes while being imperceptible to humans.
• AI Recognition Toggle (User Control): Users can conceptually choose whether their image output is "AI-Friendly" or "Protected".
• Human-Friendly Output: The protection is designed to be invisible, with no visible watermarks or distortions.
• Screenshot Resilience (Conceptual): The protective noise is embedded within the pixel data, intended to remain effective even if screenshotted.

The "Noise Adaptive Algorithm" achieves its protection through a unique, imperceptible "Adaptive Spider Web" structure. This is not random noise but a dynamically generated, multi-layered mesh that is centered on detected human subjects, adapts to its position, and is designed to make AI-generated outputs fail or become illogical.

For a more in-depth explanation of the mechanism, architecture, and the research behind this concept, please see the following documents:

• Chapter 1: Defense Mechanism Details
• Chapter 2: Verification Application Concept
• Chapter 3: Research & Development Journal

This project and its concepts are publicly disclosed on GitHub by Rijal Saepuloh as of June 2025. The primary purpose of this public disclosure is to establish prior art and to share these ideas with the broader community. This is done with the intent to foster discussion, further research, and potential development towards responsible AI.

This project is a contribution from an independent thinker with limited resources. If you are a researcher, developer, ethicist, policymaker, or an organization interested in discussing or collaborating on these concepts, please feel free to reach out.

Creator: Rijal Saepuloh Email: [email protected]