Face deepfake technology has matured, and related cases have increased year by year, with a wide range of impacts. However, with the rise of new AIGC (such as language model text generation, AI drawing, etc.) technologies, deepfake technology will be upgraded again and expanded to more levels. These new technologies are likely to generate more criminal cases. Therefore, we will sort out the context of deepfake technology from the past to the future. Our content will cover deepfake technology demonstration, real case study, technical introduction, and introduction of deepfake identification technology.

Why 5G Security = (3GPP + IT + IoT + CT) Security? To answer this question, we break it down to several questions: What is inside a cellular network? How do mobile phones connect? What is ORAN? What is 5GC? What is MEC? What is private 5G? Why traditional IT firewalls aren't enough? We introduce the architecture of Cellular Networks and discuss how it differs from traditional IT. Then we cover private 5G systems and their unique security challenges, with examples. To wrap up, we discuss a Zero-Day, and show how the private, isolated subnet of a manufacturing plant is thrown open to attackers by the peculiarities of CT networks.

As the defending side, blue teams have many different perspectives when it comes to responding to cybersecurity incidents compared to the red team. Besides considering the analysis's accuracy under a large amount of log data, blue teams also have more emphasis on response time and costs due to the needs of complying with government policies and regulations.

CyCraft has conducted extensive research on using Machine Learning techniques in investigations of large-scale cybersecurity incidents, and has incorporated AI into various scenarios, including endpoint event correlation, CmdGPT analysis of Cmdline, and automatic forensic report generation. Since last year, we have incorporated a new generation of large language models into our blue team AI assistant for automatic incident response and issue management. In this talk, the speaker will share our practical experiences and exclusive insights on how cybersecurity teams can properly use AI, and share some actual case studies.

Modern information security’s last line of defense depends on the confidentiality of devices. When the hardware implements encryption and decryption operations, it is easy to inadvertently leak sensitive information and affect security, such as execution time, power consumption, electromagnetic radiation, etc. This type of research was called Side-Channel Cryptanalysis more than 20 years ago, but it is now called Side-Channel Analysis (SCA) and has received widespread international attention. The speech will mention that compared with traditional analysis methods, such as Simple Power Analysis (SPA), Correlation Power Analysis (CPA), Template Attack, etc. The application of Machine Learning to enhance the side-channel analysis technology, regardless of whether the encryption and decryption algorithm is protected, the analysis results based on Deep Learning have been proven to have an overwhelming advantage in recent years.

As artificial intelligence (AI) has been widely adopted in a variety of applications, it is time to consider the security and privacy of AI. For example, we may wonder whether the AI model will be backdoored such that the face recognition system for access control is useless. We may also have a question about whether object detectors will be fooled such that the self-driving cars cause the accidents. We might also wonder whether the chatGPT will output illicit or discriminative contents. In this talk, we assume that the audience has certain background on AI and summarize the security and privacy issues of AI. 

Due to the explosion of applications in 5G, AIoT, and Industry 4.0, the transformation towards computerization, digitization, and intelligence has led to a continuous occurrence of supply chain attacks, making supply chain cybersecurity a globally concerning issue in recent years. Open Source Software has become a mainstream trend in the information and communications field in the past few years, and the industry frequently uses open source software as a component for integration or secondary development. Therefore, building an open source secure supply chain is a crucial issue. Since building an open source secure supply chain requires collaboration between the upstream and downstream industries, SZ will share how major international companies and different industries use process management regulations to identify the entry and exit points that should exist in organizational processes, policies, or training. Combined with Software Bill of Material (SBOM) to identify and track compliance and security status of software used and actually deployed, trust can be established among different companies to exchange deliverables composed of open source software. In addition to sharing industry examples, SZ will also share the future development blueprints and first-hand information from the OpenChain open source security standards and the international SBOM standard SPDX, hoping to build an open source secure supply chain together.