Schedule – Oct 22, 2024

The AI Revolution

Download the schedule as a PDF file: Schedule
To see the abstract of a talk click on its title.

Talks with a gray background are talks related to the main topic “The AI Revolution”.

Expand All +

  • Arena


  • In this talk, Daniel covers how defend your Security Program from AI consultancies. In other words, how to use AI to help run your program—before someone else does. It goes over how to build an AI-based structure for understanding your Security Program, and how to use AI to manage said program over time. Attendees will leave with not only a new way of thinking about AI, but a new way of thinking about security programs as well.
    The AI Revolution

  • The current unstable geopolitical situation has once again, after many years, brought back the everlasting threat of a nuclear conflict.

    This talk is intended to contribute further to a better understanding of the different cyber-physical attacks that may impact nuclear facilities, driven from the perspective of hypothetical, but realistic, state-sponsored operations. One of the novel aspects of this research is that it is based on the analysis of specific digital Instrumentation and Control equipment (Class 1E), that is currently deployed in multiple Nuclear Power Plants across Europe (including Swiss NPP such as Beznau or Gösgen), USA, Russia, or China.

  • Technology and AI have the potential to reshape propaganda and disinformation tactics, enabling new capabilities in both creation and dissemination. I explore how modern propaganda operations may benefit from content generation technologies like large language models to spread disinformation broadly. I will discuss the use of deepfakes, AI-fueled personas, and how they may be employed by state and non-state actors to influence public opinion, undermine institutions, fuel information warfare, or enhance scams. Additionally, I will introduce a systematic approach to analyzing influence campaign content to classify these threats, a process that may also be enhanced by AI/LLMs.
    The AI Revolution

  • Cyber insurance polarizes. Some view it as the one measure that will keep their company afloat after a cyber incident, others are convinced that it will not pay out in any case. Who is right?

    In our discussion we will cover what cyber insurance is, what it is not and where the big open questions are.

  • Large language models can be manipulated through language - and that means: Social engineering works with chatbots! This is good news, because it helps us to use large language models for our purposes (and possibly differently than they are intended). But at the same time this is bad news, because the bad actors can also do this. The talk uses examples from my recent research to explain how large language models can be manipulated. I show how I got them to reveal their dark secrets - like manipulative initial prompts - and thus exposed the companies behind them and their shady activities. Or how they helped me with investigative research, developed and explained the best Google Dorks, removed redactions and revealed things that they are not supposed to reveal. This is a lot of fun. But it also shows: LLMs will always leak our data, they can be manipulated and they will always say things they are not supposed to say.
    The AI Revolution

  • AI Security has been researched for almost two decades. Yet, existing, frequently studied threat models have never been tested in terms of real-world usage of AI. In this talk, we discuss a survey with 271 real-world AI practitioners, whose description of AI usage we match with existing threat models. While we find that all threat models do exist, there are also significant mismatches where research is too generous with the attacker.
    The AI Revolution

  • Your metrics are boring and dangerous. Recycled slides with meaningless counts of alerts, incidents, true and false positives… SNOOZE. Even worse, it’s motivating your team to distort the truth and subvert progress. This talk is your wake-up call to rethink your detection and response metrics.

    Metrics tell a story. But before we can describe the effectiveness of our capabilities, our audience first needs to grasp what modern detection and response is and its value. So, how do we tell that story, especially to leadership with a limited amount of time?

    Measurements help us get results. But if you’re advocating for faster response times, you might be encouraging your team to make hasty decisions that lead to increased risk. So, how do we find a set of measurements, both qualitative and quantitative, that incentivizes progress and serves as a north star to modern detection and response?

    Metrics help shape decisions. But legacy methods of evaluating and reporting are preventing you from getting the support and funding you need to succeed. At the end of this talk, you’ll walk away with a practical framework for developing your own metrics, a new maturity model for measuring detection and response capabilities, data gathering techniques that tell a convincing story using micro-purple testing, and lots of visual examples of metrics that won’t put your audience to sleep.

  • Cybersecurity is on the verge of a radical transformation, driven by the rise of Generative AI (GenAI). Within the next five years, human-led penetration testing will give way to AI-powered solutions, and offensive security actions will be teleoperated or fully automated. GenAI companies are already undermining the security capabilities in their models to avoid public backlash, leaving critical gaps in defense.* In this talk, Víctor Mayoral Vilches will present his groundbreaking work on the "Levels of Autonomy in Cybersecurity," a framework that maps the shift from human-led pentesting to fully autonomous Cybersecurity AIs. He begins with PentestGPT, an open-source AI tool launched in 2023 that assists with offensive security planning and sparked widespread controversy. From there, Víctor will explore semi-automated tools before introducing the Robot Immune System (RIS), a cutting-edge, AI-driven security solution. RIS employs Artificial Intelligence and Game Theory to autonomously understand, predict, and mitigate cyber threats in real time, evolving dynamically to meet the ever-changing cybersecurity landscape. Originally developed for robotics, RIS has expanded its reach to IT and OT environments, fueled by the latest advances in GenAI. Víctor will conclude with insights into his team’s current research, which is focused on advancing these Cybersecurity AIs to comply with the European Union's NIS2 and AI Act, paving the way for a future where AI not only defends but outsmarts cyber adversaries.
    The AI Revolution

  • John Graham-Cumming, CTO of Cloudflare, discusses the evolution of web application firewalls (WAFs) as a model for AI in general. He has championed the integration of machine learning into WAFs, and the challenges of adapting to new and evolving web attacks. His presentation explores the history of spam filtering as an analogy for the potential of machine learning in enhancing WAF effectiveness, while also cautioning against the potential for attackers to use machine learning for evasion: This is the beginning of a new era, the cat-and-mouse game is changing, but it's far from over.
    The AI Revolution

  • Our standing dinner allows everybody to mingle, meet friends and talk about the many interesting talks of the day. Naturally, most of the speakers will still be around, so don't rush off after the last talk.
    break-slot

  • Distribution of the prizes of the Sponsoring Raffle.

  • Scenario


  • You have an AI project, but are stuck with getting legal approval? David will share some insights on how to deal with lawyers and legal departments with regard to AI. What are the aspects they are looking for? What helps you getting the green light? How to do in terms of governance?
    The AI Revolution

  • Language models are currently near the peak of the hype curve. Their application to cybersecurity has been a topic of academic research for a while. In this talk we present the results of our efforts to put one of the many proposed architectures into production. We explain how and where AI can fit into security systems and detail the approach we took. We also elaborate on the problems we faced and detail why there is a big gap between what researchers put out and what is feasible and useful in practice.
    The AI Revolution

  • As we try to understand LLMs and AI in general, navigating its potential benefits and the myriad of security challenges it presents, one fundamental question emerges: What do we, both as individuals and as democratic societies, truly desire? More crucially, how can we collaborate to fortify our digital ecosystems, ensuring they not only realize our aspirations but also safeguard our democratic values, human rights and the integrity of our society? This talk introduces the current political initiative of the Pirate Party Zürich titled "For A Fundamental Right To Digital Integrity", which seeks to address these core concerns by advocating for a legal framework that not only promotes safety and human-centric AI but also underpins our collective security.
    The AI Revolution

  • Traditional security strategies often face resistance. This is not due to human failings. It is due to a lack of understanding of the socio-dynamics at play within organizations and human beings. This talk explores how purpose-driven security can overcome resistance. It can transform security efforts from a source of friction to a wellspring of action. We can unlock a powerful force for change. True motivation comes from a sense of purpose and shared goals. When people see how their work fits into the big picture, they want to help protect the company's future. They understand that security safeguards that vision. You'll learn, through real world examples and a step-by-step guide, to understand and use socio-dynamics. You can create a purpose-driven security culture.

  • APIs are a foundational technology in today’s app-driven world and increasingly becoming the main target for attackers. How do you protect yourself? This talk will walk you through the techniques attackers use against APIs like broken object level authorization (BOLA) by following a typical API pen testing methodology. For each phase and attack, the tables are turned by covering how the attack looks from the defender's point of view including proactive ways to catch attacks early. You’ll understand how attackers find and exploit vulnerabilities and gain insight into why many traditional AppSec approaches fall short for APIs. The goal is to provide a complete overview of API vulnerabilities from both attack and defense perspectives so you can ramp up your testing and protection of all the new APIs in your AppSec life.

  • Let's rethink our approach to human factors in information security amidst emerging AI threats. This session advocates a human-centred approach, placing people and processes at the core of security design.

    Common pitfalls in managing human risk will be highlighted, particularly with the rise of AI-driven hacking techniques. Current practices will be examined through the latest research, focusing on the evolving threat landscape.
    The AI Revolution

  • This talk will explore the key developmental phases of building a successful Security Champions Program and how it can transform into a security powerhouse. We’ll cover how to identify the right candidates, structure the program, foster engagement and growth, and use data-driven strategies to gain leadership buy-in. Attendees will gain practical insights into creating a Security Champions Program that strengthens security culture across the organization.

  • Our standing dinner allows everybody to mingle, meet friends and talk about the many interesting talks of the day. Naturally, most of the speakers will still be around, so don't rush off after the last talk.
    break-slot

  • Distribution of the prizes of the Sponsoring Raffle.

  • Sponsoring - Panorama


  • To protect against the increasing frequency and sophistication of cybercrime, organizations are deploying a variety of security solutions. This increases the complexity and scale of the security landscape, especially as today both environments (on-premises and multi-cloud) need to be protected. Learn what BAS is and how it helps improve cybersecurity, how it is implemented at the customer site, and how BAS supports the incident response process.

  • Since generative AI became available to the public, there has been a sharp rise in successful phishing campaigns and ransomware attacks. Cybercrime is becoming even faster, more automated and more professional. In one recent case, encryption occurred within just 8 seconds of initial access. However, machine learning and deep learning are also empowering defenders. These technologies are enabling early detection and mitigation, significantly enhancing cybersecurity solutions. In this talk, I will share striking insights from real-world cyberattacks, discuss the recent work of our CSIRT, and explore the evolving landscape of threat actors. Finally, I will delve into current challenges, trends, and the growing role of AI in cyber defense.
    The AI Revolution

  • Most organizations are in the process of preparing for a high impact incident. Our focus as a security community is now on developing the right processes, having the right tools, and getting everyone involved.

    Looking back over seven years of responding to major security incidents, I keep coming back to the same observation: We need to make sure that we keep the human side of incidents in mind. We must ensure that we prepare and act in a way that considers the people involved in responding to an incident, those affected by an incident and their wider communities. If we don't, we will miss and fail to address an important impact.

    In this talk, I invite you to challenge your incident response management system to ensure that the human side of the response is as well prepared as all the other important elements.

  • Although the TPM sniffing attack has been known and utilized for nearly five years, it has evolved significantly since its introduction. The attack process has become simpler and more generalized, resulting in increased speed and reliability. This presentation will first share insights gained from using this attack on dozens of machines over the years, highlighting key factors that can enhance its execution. The reduction in complexity and attack time now makes it possible to compromise a machine in just a few minutes, infect it, and then restore it before the owner even realizes it was gone. Secondly, the presentation will explore the multi-factor authentication provided by BitLocker. While multi-factor authentication is crucial for maintaining a baseline level of security, various attack scenarios remain possible even with this configuration. For instance, a malicious user could use TPM sniffing to escalate privileges on a machine if they know the second authentication factor. Although this possibility has been discussed in several publications, Microsoft's documentation on BitLocker is only partial, and some mechanisms remain unexplored. No existing tool has been able to execute this attack when BitLocker is not in transparent mode. This part of the presentation will delve into the operating system’s inner workings, examining the Windows bootloader and its interactions with the TPM. The goal is to understand how multi-factor authentication works and how it might be bypassed, enabling to decrypt the disk and gain high-privileged access to the operating system.

  • AI poses data security, compliance, and privacy challenges for organizations that, if not addressed properly, can slow down adoption of the technology. Due to a lack of visibility and controls to protect data in AI, organizations are pausing or in some instances even banning the use of AI out of abundance of caution. To prevent business critical data being compromised and to safeguard their competitive edge, reputation, and customer loyalty, organizations need integrated data security and compliance solutions to safely and confidently adopt AI technologies and keep their most important asset – their data – safe. Michael and Umberto are excited to share how Microsoft significantly invests into cybersecurity and help empower customers to protect and govern their data. We approach data protection in various ways, and Microsoft Purview plays a key role that provides a set of comprehensive, innovation driving solutions – addressing data security, compliance and privacy.
    The AI Revolution

  • Bruno Blumenthal will provide an overview of the critical frameworks and upcoming regulations, shaping AI governance. This presentation highlights key compliance requirements, including ISO standards, FINMA regulations, and the EU AI Act. Attendees will gain essential insights into current standards and emerging regulatory trends, helping them to navigate the complex AI compliance landscape effectively.
    The AI Revolution

  • Our standing dinner allows everybody to mingle, meet friends and talk about the many interesting talks of the day. Naturally, most of the speakers will still be around, so don't rush off after the last talk.
    break-slot

  • Distribution of the prizes of the Sponsoring Raffle.