[DSC MENA 24] Ramy_Agieb_-_Advancements_in_Artificial_Intelligence_for_Cybersecurity-Ramy_Agieb_19-4-V1.pptx

With the continuous growth of the digital environment, the risks in the online realm also increase. This calls for strong security measures to safeguard valuable information and essential systems. Artificial Intelligence (AI) has become a powerful weapon in the fight against cyber threats. This talk presents a thorough examination of the most recent algorithms and applications of artificial intelligence in the field of cybersecurity.

1. 2 ?????????????????????? From “Wireless Communications” Edfors, Molisch, Tufvesson d Advancements in Artificial Intelligence for Cybersecurity

2. 3 From “Wireless Communications” Edfors, Molisch, Tufvesson d Hackers using AI tools

3. 4 WormGPT From “Wireless Communications” Edfors, Molisch, Tufvesson d A new generative AI cybercrime tool Underground forums as a perfect tool for adversaries to launch sophisticated phishing campaigns and business email compromise (BEC) attacks . BEC cost 1.8 billion in 2020 For Learning only

4. 5 TorGpt From “Wireless Communications” Edfors, Molisch, Tufvesson d For Learning only

5. 6 ????????

6. 7 Artificial Intelligence From “Wireless Communications” Edfors, Molisch, Tufvesson “Can a machine think and behave like humans do?” According to the father of Artificial Intelligence John McCarthy , it is “ The science and engineering of making intelligent machines , especially intelligent computer programs”.

7. 8 Agenda d

8. 9 Agenda d

9. 10 Artificial Intelligence From “Wireless Communications” Edfors, Molisch, Tufvesson d

10. 11 AI & ML & DL & FM From “Wireless Communications” Edfors, Molisch, Tufvesson d AI is the broader concept of creating intelligent machines . ML is a subset of AI focusing on learning from data . DL is a subset of ML employing deep neural networks for learning and modeling complex patterns . FM Generative AI , any model that is trained on broad data (generally using self-supervision at scale )

11. 12 AI & ML & DL & FM From “Wireless Communications” Edfors, Molisch, Tufvesson d

12. 13 AI Initialization From “Wireless Communications” Edfors, Molisch, Tufvesson d AI Initialization

13. 14 M easure the Performance d Confusion Matrix Popular evaluation metric used to describe the performance of a classification model (or "classifier"). FP: lead to unnecessary alerts and operational., consuming valuable resources and causing alert fatigue FN: can result in undetected security breaches .

14. 15 AI & ML & DL From “Wireless Communications” Edfors, Molisch, Tufvesson d Handle False Positives False Negative

15. 16 Agenda d

16. 17 Cybersecurity From “Wireless Communications” Edfors, Molisch, Tufvesson d Cybersecurity is the practice of protecting smart divices like Mobile, computer systems, networks, and data from unauthorized access, attacks, damage, or theft .

17. 18 Core Principles of Cybersecurity Confidentiality : Ensuring that sensitive information is only accessible to authorized individuals or systems. Integrity : Protecting data and information from being altered without authorization . Availability : Ensuring that systems and data are available and accessible when needed. CIA

18. 19 Agenda d

19. 20 AI & Cybersecurity From “Wireless Communications” Edfors, Molisch, Tufvesson d AI providing new ways to detect and prevent cyber threats in real time, and digital forensics. Set of challenges, including issues of data privacy, false positives, and the need for constant updates and adaptations.

20. 21 AI & Cybersecurity From “Wireless Communications” Edfors, Molisch, Tufvesson d Some key features of AI technology in cybersecurity include: Real-time monitoring: AI continuously monitor network sand systems for suspicious activity, identifying and reacting to potential threats in real time. Behavior analytics: AI analyze user behavior and detect anomalies that may indicate a security breach or attempted attack. Anomaly detection: AI can identify patterns that may indicate a cyberattack , even in cases where the attack may not fit a predefined threat profile .

21. 22 AI & Cybersecurity d Real-time monitoring Behavior analytics Anomaly detection

22. 23 AI & Cybersecurity From “Wireless Communications” Edfors, Molisch, Tufvesson d

23. 24 Alogorithmes of A I in Cybersecurity d Fully Homomorphic Encryption (FHE) Long Short-Term Memory (LSTM) Networks Variational Autoencoders (VAEs) Natural Language Processing (NLP) Models Graph Neural Networks (GNNs)

24. 25 AI & Cybersecurity From “Wireless Communications” Edfors, Molisch, Tufvesson d

25. 26 Alogorithmes of A I in Cybersecurity d Merkle Trees Secure Multi-Party Computation (SMPC) Differential Privacy Zero-Knowledge Proofs (ZKPs) Integrity

26. 27 AI & Cybersecurity From “Wireless Communications” Edfors, Molisch, Tufvesson d

27. 28 Alogorithmes of A I in Cybersecurity d Recurrent Neural Networks (RNNs ) Availability Long Short-Term Memory (LSTM) networks Isolation Forests Q-Learning Streaming Analytics Generative Adversarial Networks (GANs)

28. 29 Alogorithmes of A I in Cybersecurity d detect Deepfake Facial Landmark Detection XceptionNet FaceForensics ++ Voice Biometrics Fusion of Multiple Modalities

29. 30 AI & Cybersecurity Python Libraries From “Wireless Communications” Edfors, Molisch, Tufvesson d

30. 31 AI & Cybersecurity Python Libraries From “Wireless Communications” Edfors, Molisch, Tufvesson d

31. 32 Agenda d

32. 33 AI & Cybersecurity Tools d Top Tools Intrusion Detection Network Traffic Analysis Malware Analysis Data Preprocessing WEKA Waikato Environment for Knowledge Analysis : open-source software package with a wide range of tools and algorithms for data preprocessing , machine learning , and data mining tasks Log Analysis Threat Intelligence Analysis

33. 34 AI & Cybersecurity Tools From “Wireless Communications” Edfors, Molisch, Tufvesson d Top Tools AI platform designed for advanced cybersecurity and penetration testing applications Threat Detection and Prevention Vulnerability Assessment Custom AI Solutions Incident Response

34. 35 AI & Cybersecurity Tools From “Wireless Communications” Edfors, Molisch, Tufvesson d Top Tools

35. 36 AI & Cybersecurity Tools using by HACKERS From “Wireless Communications” Edfors, Molisch, Tufvesson d

36. 37 Agenda d

37. 38 AI & Cybersecurity From “Wireless Communications” Edfors, Molisch, Tufvesson d Ethical Considerations in AI-Cybersecurity Integration Bias: The data used to train AI models is representative , diverse , and unbiased . Transparency: Ensuring that AI systems are explainable and understandable can help prevent suspicion and mistrust. Accountability: It is important to clearly define who is responsible for errors or misuse of AI technology.

38. 39 AI & Cybersecurity d The Role of Government Agencies In the United States , the Department of Home land Security (DHS) Science and Technology Directorate has launched several initiatives aimed at improving cybersecurity through AI . These initiatives include cybersecurity research and development , technology testing and evaluation , and technology transition and commercialization.

39. 40 AI & Cybersecurity d The Role of Government Agencies National Institute of Standards and Technology ( NIST ) developed a frame work for improving critical infrastructure cybersecurity. National Security Agency ( NSA ) has created a Cybersecurity Directorate to enhance the agency’s cyber defense capabilities. Academic institutions are conducting cutting-edge research in the field of AI and cybersecurity.

40. 41 From “Wireless Communications” Edfors, Molisch, Tufvesson d AI technology is playing a critical role in safeguarding sensitive information from data breaches, theft, and manipulation .

41. 42 From “Wireless Communications” Edfors, Molisch, Tufvesson d