Wednesday, May 15: This workshop highlights the threat landscape in AI systems and the concept of adversarial attacks. It builds on the previous workshop and presents real-world examples to emphasize the importance of addressing these threats. The key concepts covered include threat models, adversarial attacks, model robustness, and defenses.

Wednesday, May 29: This workshop provides a broad understanding of adversarial attacks in AI by focusing on the different types of attack, their methods, and their potential impact on AI systems. The workshop delves into different types of adversarial attacks, such as evasion, poisoning, and model inversion under various threat models (e.g., white- and black-box settings). Moreover, it addresses aspects such as transferability, single-class and targeted attacks.

Wednesday, June 12: This workshop provides a comprehensive understanding of the key concepts related to robustness and resilience in AI models. These concepts include general and adversarial robustness (i.e., the ability of the model to perform consistently and accurately in the presence of uncertainties and perturbations), domain adaptation and transfer learning, model generalization, and fault tolerance. These concepts help participants understand the challenges facing AI systems and learn strategies to improve their stability, reliability, and performance.

Wednesday, June 26: This workshop introduces key concepts of privacy in the context of AI and explores privacy-preserving methods around differential privacy and federated learning. The workshop discusses challenges and methods to preserve privacy while leveraging sensitive data for AI model training. This includes understanding potential privacy risks caused by either unintended memorization of sensitive data or adversarial means (e.g., model inversion and inference attacks). The discussion also includes describing the privacy budget and trade-offs between privacy and utility. The panel will discuss the benefits, challenges, and potential risks in federated learning and decentralized data settings.

Wednesday, July 10: The goal of this workshop is to explore ethical considerations in AI, specifically focusing on the concepts of bias and fairness. Participants will gain an understanding of the challenges associated with bias in AI systems and the importance of promoting fairness and mitigating discrimination. The panel discussion will focus on the responsibility of AI practitioners to address bias (i.e., different types of biases such as algorithmic bias, data bias, and societal bias) and to prompt fairness in AI systems. Considering real-world examples, e.g., a biased hiring system, studying the impact of bias and approaches to mitigate bias, including data pre-processing techniques, algorithmic adjustments, and fairness-aware model training.

Wednesday, July 24: This workshop explores the importance of trust in AI systems and introduces participants to the concepts of transparency, explainability, and interpretability. Participants will understand the role of these concepts in building trust, ensuring accountability, and addressing the black-box nature of AI models. The key concepts addressed in this workshop are, 1) trust in AI and its significance in fostering user and societal acceptance, 2) explainability and interpretability in AI and the ability to understand and trace the decision-making process of AI models, and 3) techniques for explainability and interoperability including examples for post-hoc techniques and inherently interpretable models.

Wednesday, August 7: This workshop introduces the concepts of DevOps (Development and Operations) and MLOps (Machine Learning Operations) in the context of AI security. Participants will understand the importance of integrating security practices into the AI development lifecycle and learn how DevOps and MLOps can enhance the security and reliability of AI systems. The key concepts covered in this workshop are: 1) DevOps practices in the context of AI development to improve collaboration, efficiency, and quality, 2) Integration of security into the DevOps workflow, 3) MLOps practices and tools to manage the lifecycle of machine learning models, including development, deployment, and monitoring, 4) Security considerations during the model development phase, 5) Regulatory requirements and industry standards, such as GDPR, HIPAA, ISO 27001, and NIST Cybersecurity Framework.

Wednesday, August 21: This workshop introduces key regulations and standards related to AI and data governance. Participants will understand the legal and ethical considerations involved in AI development and deployment and the importance of compliance with regulations and adherence to industry standards. The concepts covered are: 1) Introduction to data governance in the context of AI, including data collection, storage, processing, and sharing, and the importance of ensuring the privacy, security, and ethical use of data in AI systems. 2) Introduction to legal (e.g., GDPR, CCPA, and the EU AI Act, and ethical frameworks around AI systems. 3) Data privacy and protection principles, e.g., consent, purpose limitation, and the right to erasure, and methods for data protection, e.g., anonymization, pseudonymization, and encryption. 4) Responsible AI and impact assessment.

Wednesday, September 4: This workshop highlights the importance of secure deployment and operation of AI systems and introduces key concepts and best practices to ensure the security and resilience of AI systems throughout their lifecycle. The emphasis in this workshop is on the secure deployment of AI models and the importance of collaboration between development, operations, and security teams during the operational stage of the system lifecycle to maintain system integrity.

Wednesday, September 18: This workshop analyzes real-world cases and outline practices and challenges with respect to cybersecurity and privacy in AI systems.

Wednesday, October 2: This workshop analyzes real-world cases and outline practices and challenges with respect to cybersecurity and privacy in AI systems.