About the Technology

False + generates privacy-first art and apparel by leveraging best practices in modern image generation and machine learning. Our adversarial patterns are crafted using advanced algorithms—such as neural network-based image synthesis and adversarial perturbation generation—rooted in peer-reviewed research and the latest findings in computer vision.

Our creative and technical process is inspired by rigorous, reproducible research in adversarial machine learning and synthetic image creation. Patterns are tested against a variety of AI recognition models to maximize their capacity to confuse, camouflage, or disrupt automated identification.

Research Foundations and Best Practices

• Adversarial Example Generation:
  Following methods introduced in “Explaining and Harnessing Adversarial Examples” (Goodfellow et al.), we use gradient-based optimization to generate perturbations that specifically target vision models.
• Physical-World Adversarial Attacks:
  Our approach is influenced by works like “Robust Physical Perturbations Against Real World Object Detectors” (Eykholt et al.) and Kurakin et al., "Adversarial Examples in the Physical World", adapting digital attacks for real-life wearable art.
• Generative Models and Neural Synthesis:
  Our workflows also build on leading advances in image generation, including GANs as described in “Generative Adversarial Networks” (Goodfellow et al.), and the neural art generation techniques outlined in Gatys et al., "A Neural Algorithm of Artistic Style".
• Ongoing Research & Verification:
  We refine our patterns based on fresh findings and datasets—see Adversarial Machine Learning Resources—and aim for reproducibility and empirical robustness in our art and apparel.

Our goal is to blend rigorous technical research with creative expression, offering pieces that challenge the boundaries of surveillance, art, and technology. While no method is absolute, our work follows the highest standards currently recognized in the adversarial machine learning and generative art communities.

Disclaimer: No adversarial technique or artwork can guarantee protection against all AI surveillance systems. Effectiveness varies; wear and display responsibly, and stay informed.

Legal information