Lec01[1].pdf

10/9/2024 AI in Security Issue Dr.Mohamed Moustafa Associate Professor , Computers and Artificial Intelligence CIO – DMU,SAS, MS 1 Course Learning Objectives Understa...

10/9/2024 AI in Security Issue Dr.Mohamed Moustafa Associate Professor , Computers and Artificial Intelligence CIO – DMU,SAS, MS 1 Course Learning Objectives Understand cybersecurity challenges and the role of AI in addressing them. Identify and evaluate vulnerabilities in AI systems, including trapdoors and adversarial attacks. Explore AI's implications in offensive and defensive cybersecurity measures. Gain hands-on experience in AI- driven security applications and threat simulations. 2 1  10/9/2024 Class Rules You can do anything except: Make noises (chatting, singing…) Feel free to interrupt me if you have questions. According to the university policy, taking attendance is needed. Important: you are required to have an 80% attendance to be able to seat for the final exam. 3 3 Course Assessment  Temporary according to the situation:  Final exam:50%  Midterm: 20%,  Quizzes: 10%  Project: 20%, 2-3 members per group; report and presentation are required.  Important:cheating and plagiarism will get no marks. 4 4 2  10/9/2024 A few suggestions…. Your final grade is based on points – not on an accumulation of grades. You start the class with zero points and earn your way to your final grade If you have an issue or problem, communicate – send me an email If you know you’re not going to meet the deadline for a quiz or assignment – email me BEFORE the deadline 5 5 Introduction to AI and Security AI Security Issues 6 3  10/9/2024 Learning Objectives Understand the basic concepts of AI and its applications. Recognize the increasing importance of AI in cybersecurity. Identify key AI security challenges and potential threats posed by AI technologies. Understand how AI enhances both offensive and defensive cybersecurity measures. 7 Data challenges … Agenda “Data Everywhere, You have data , You have Everything ” 8 4  10/9/2024 DATA INTELLIGENCE 9 10 5  10/9/2024 ARTIFICIAL INTELLIGENCE 1. © 2020 GEOCODE 11 What is AI? AI refers to systems that perform tasks requiring human intelligence (e.g., decision-making, pattern recognition). 12 6  10/9/2024 Categories of AI: Narrow AI: Performs specific tasks like spam filtering. General AI: Hypothetical AI with human-level intelligence. 13 What is Machine Learning (ML)? A subset of AI that trains systems to learn from data without being explicitly programmed. 14 7  10/9/2024 Machine learning 15 Machine Learning Supervised Learning: Model is trained on labeled data (e.g., malware detection). Unsupervised Learning: Model identifies patterns without labels (e.g., anomaly detection). Reinforcement Learning: The system learns by interacting with its environment and receiving feedback (e.g., AI in autonomous cybersecurity systems). 16 8  10/9/2024 AI Enabling Security AI systems are increasingly used to support security. Can detect anomalies, predict potential attacks, and help secure systems against known and unknown threats. Example : – AI in Intrusion Detection Systems (IDS) – AI-driven Malware Analysis 17 Security Enabling Better AI Example : – Homomorphic Encryption allows computations on encrypted data without decrypting it, ensuring privacy in AI models that handle sensitive data like health records. – Federated Learning lets AI models be trained across multiple devices or servers while keeping the data decentralized. This approach enhances privacy by avoiding direct access to the data itself. 18 9  10/9/2024 Why AI Security is Important AI is becoming embedded in various sectors of critical infrastructure, from smart cities managing traffic flow to healthcare systems diagnosing diseases and predicting patient outcomes. As AI increasingly takes over decision- making processes, it's critical to understand that attackers are more incentivized to target these systems due to the potential damage. This understanding is key to being prepared for the evolving security landscape. Example: hacking an AI-powered medical device could result in incorrect diagnoses or even harm to patients. 19 Role of AI in Cybersecurity – Defense: Detects and prevents cyber threats through pattern recognition and real-time analysis. – Offense: Used by attackers to automate attacks and create sophisticated malware. 20 10  10/9/2024 Why AI is critical in modern cybersecurity? Automation Speed and Accuracy Adaptive Defense 21 AI Applications in Cybersecurity Intrusion Detection Systems (IDS) using anomaly detection. Malware Detection with machine learning models. Behavioral Analysis for insider threat detection. 22 11  10/9/2024 Challenges of AI in Security AI as a Vulnerability: Example: – Adversarial Machine Learning (AML) Attackers can craft inputs that cause AI systems to make mistakes 23 AI Vulnerabilities Data Poisoning: Manipulating training data to degrade performance. Adversarial Attacks: Introducing perturbations to fool AI models. Model Inversion Attacks: Extracting sensitive data from AI models. 24 12  10/9/2024 AI Vulnerabilities A company uses AI for autonomous threat detection. The "Attackers" must come up with ways to exploit the AI system, and the "Defenders" must create strategies to prevent this. 25 Trust and Bias Issues AI Bias Lack of Transparency 26 13  10/9/2024 The Dual Use of AI Offensive AI: Attackers leverage AI for smart malware, spear-phishing, and creating deepfakes for fraud. Defensive AI: AI is deployed to predict, detect, and respond to security threats before they can cause harm. 27 The Importance of AI in Modern Cybersecurity Scalability and Efficiency Evolution of Cyber Threats 28 14  10/9/2024 Examples of AI in Cybersecurity Tools Darktrace IBM Watson for Cybersecurity Vectra AI 29 15