{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"Embracing Digital Transformation","title":"#205 GenAI and Cybersecurity","html":"","width":"100%","height":180,"duration":1865,"description":"In this episode, Darren interviews returning guest Steve Orrin, CTO of Intel Federal. about the intersection of Artificial Intelligence (AI) and cybersecurity. Embracing AI's potential for bolstering cybersecurity while ensuring the security of AI itself requires a balance that necessitates early preparation and innovative strategies.Amidst the ever-evolving world of technology, the convergence of Artificial Intelligence (AI) and cybersecurity has sparked a compelling discourse. Today, we delve into insights from a thought-provoking conversation with Steve Orrin, the esteemed CTO of Intel Federal. We explore the security implications of AI and the innovative strides being made to establish a secure AI ecosystem. Securing the AIIn the realm of AI, the paramount task is to secure the solution and its pipeline. The dynamic nature of AI necessitates a broader scope of surveillance, extending beyond the management of static applications. This vigilance encompasses data sources, evolving models, and weight changes that influence AI outcomes, presenting a formidable challenge.The security struggles are further aggravated by unreliable data input from various sources. Conventional cybersecurity techniques have proven to be inadequate when fighting against AI manipulation and interference. Given this complexity, the concept of continuous testing and validation of AI emerges as a plausible solution. The constant testing approach helps identify manipulative instances in the AI's learning process by regularly monitoring the ‘confidence levels’ and aiding the reinforcement of its original training.  Bringing Pre-Deployment Strategy Post-Deployment The concept of continuous validation presents a challenging perspective. As per the DevSecOps rule, isolation and separation of environments are essential. However, constant development is the norm in AI, making it almost necessary to import pre-deployment testing methods into post-deployment stages. This approach births the idea of...","thumbnail_url":"https://img.transistorcdn.com/IRrW2aizIeoZDn3gKLEax-JYQ8V_WzaFpHdgsslDx3k/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9jM2Ji/MDk1OTdiYzA4ZWMw/NWNlOTY0N2RhMWQ3/YmY5Mi5wbmc.webp","thumbnail_width":300,"thumbnail_height":300}