Alyssa Miller has always enjoyed breaking things to see how they work and how to fix them. She is a security professional, hacker, researcher, speaker and cyber security evangelist with over 15 years of experience in the security industry. Her professional career began as a hacker performing various security assessments for a large financial services firm. For the last 7 years, Alyssa has expanded her expertise working as a consultant for multiple US based security firms. Over the course of this work Alyssa has been working with senior and executive leadership in organizations ranging from government agencies and small business to the top Fortune 500 corporations to develop comprehensive and effective cyber security programs. Currently, Alyssa heads the Information Security Solutions practice for Illinois-based CDW Corporation.
Talk: The Death of Trust: Exploring the Deepfake Threat
Deepfake media is becoming more common and easy to produce and it is expected to have a major impact on public discourse in upcoming elections. The use of Generative Adversarial Networks (GANs) has enabled the creation of incredibly lifelike artificial media. As a result, researchers and commercial organizations are scrambling to develop detection capabilities that can identify even the most convincing fakes.
In this session, we’ll describe the technical application of GANs when generating this media and highlight some of the most recent advances. We’ll also examine how the current limitations of this technology can be exploited by both low-tech and high order machine learning capabilities to detect deepfake media. Building upon limited published research, we’ll analyze the relative effectiveness of these techniques and explore how creators of false content could adapt in order to evade detection.
Additionally, we’ll consider those situations when detection may not always be enough. For instance, when the subject of a video claims it to be a fake or when there are national security concerns requiring us to confirm, with the highest degree of certainty, the veracity of the media. We will evaluate current technologies that seek to provide positive confirmation of digital media authenticity. Finally, we will propose further potential solutions and scrutinize their feasibility in helping maintain a high level of trust in digital media.