It Is What It Is... For Now.

Abstract: In recent years, generative adversarial networks (GANs) and its variants have achieved unprecedented success in image synthesis. They are widely adopted in synthesizing facial images which brings potential security concerns to humans as the fakes spread and fuel the misinformation. However, robust detectors of these AI-synthesized fake faces are still in their infancy and are not ready to fully tackle this emerging challenge. In this work, we propose a novel approach, named FakeSpotter, based on monitoring neuron behaviors to spot AI-synthesized fake faces. The studies on neuron coverage and interactions have successfully shown that they can be served as testing criteria for deep learning systems, especially under the settings of being exposed to adversarial attacks. Here, we conjecture that monitoring neuron behavior can also serve as an asset in detecting fake faces since layer-by-layer neuron activation patterns may capture more subtle features that are important for the fake detector. Experimental results on detecting four types of fake faces synthesized with the state-of-the-art GANs and evading four perturbation attacks show the effectiveness and robustness of our approach.

Abstract: Addictive products—tobacco, alcohol, gambling, and the like—have been considered legitimate regulatory targets for millennia, a tradition into which both Founding-era and modern America comfortably fits. Expressive products—newspapers, books, movies, and video games—on the other hand, have been considered essentially immune from content-based regulation, thanks to the First Amendment. But what if the content of an expressive product makes it addictive? Which tradition must give in: the ancient ability of legislatures to protect society at large from the wide-ranging impacts of addiction, or the legal shield that has generated a thriving culture of artistic independence? This Article is the first to explore the question of how the First Amendment should treat intentionally addictive speech. Social scientists indicate that certain behavioral addictions premised on compulsive use of expressive products—in particular, video games and pornography—are real dysfunctions of the brain, explainable in part by the intentional choices of developers and producers to create addictive products. And regulators are beginning to unsteadily lurch into action, without any evidence that they are taking the First Amendment into account. This Article proposes that, under current doctrine, any such regulation would need to satisfy strict scrutiny. It then argues for a departure and a recognition that intentionally addictive expression does not merit First Amendment coverage.

Abstract: In recent years, generative adversarial networks (GANs) and its variants have achieved unprecedented success in image synthesis. They are widely adopted in synthesizing facial images which brings potential security concerns to humans as the fakes spread and fuel the misinformation. However, robust detectors of these AI-synthesized fake faces are still in their infancy and are not ready to fully tackle this emerging challenge. In this work, we propose a novel approach, named FakeSpotter, based on monitoring neuron behaviors to spot AI-synthesized fake faces. The studies on neuron coverage and interactions have successfully shown that they can be served as testing criteria for deep learning systems, especially under the settings of being exposed to adversarial attacks. Here, we conjecture that monitoring neuron behavior can also serve as an asset in detecting fake faces since layer-by-layer neuron activation patterns may capture more subtle features that are important for the fake detector. Experimental results on detecting four types of fake faces synthesized with the state-of-the-art GANs and evading four perturbation attacks show the effectiveness and robustness of our approach.

Abstract: Addictive products—tobacco, alcohol, gambling, and the like—have been considered legitimate regulatory targets for millennia, a tradition into which both Founding-era and modern America comfortably fits. Expressive products—newspapers, books, movies, and video games—on the other hand, have been considered essentially immune from content-based regulation, thanks to the First Amendment. But what if the content of an expressive product makes it addictive? Which tradition must give in: the ancient ability of legislatures to protect society at large from the wide-ranging impacts of addiction, or the legal shield that has generated a thriving culture of artistic independence? This Article is the first to explore the question of how the First Amendment should treat intentionally addictive speech. Social scientists indicate that certain behavioral addictions premised on compulsive use of expressive products—in particular, video games and pornography—are real dysfunctions of the brain, explainable in part by the intentional choices of developers and producers to create addictive products. And regulators are beginning to unsteadily lurch into action, without any evidence that they are taking the First Amendment into account. This Article proposes that, under current doctrine, any such regulation would need to satisfy strict scrutiny. It then argues for a departure and a recognition that intentionally addictive expression does not merit First Amendment coverage.