A Conversation:

The current state of AI security

A fun, sharp interview that cuts quickly and thoroughly through the noise — and gets to what really matters right now

• Incumbents vs. AI-native? Is this really a question anymore? Speed wins. If you’re slow, your org will shadow-adopt AI overnight
• Two problem buckets: “gold-rush” adoption mistakes vs. AI’s weird, non-deterministic failures
• Don’t bet only on LLMs to police LLMs — combine classic security controls with ML alignment

PANEL:

Strategies for Safeguarding Individuals and Securing Data in the Era of Generative AI

Join the discussion where experts explore the evolving threat landscape of personal data amid rapid AI advancements. Discover actionable insights on managing risks, leveraging AI governance, and implementing effective technical controls to protect privacy and data integrity. Don’t miss this essential conversation for navigating the complexities of AI security!

PANEL:

AI Security: LLM Security and Abuse Cases

Join our panel discussion where experts explore the evolving landscape of threats posed by large language models (LLMs). Delve into the challenges of ensuring security in autonomous AI systems, the impact of democratized coding tools, and the importance of robust governance policies. Discover actionable strategies to mitigate risks and enhance defenses against emerging abuse cases in AI technology. Don’t miss this critical conversation!

PANEL:

AI Security: Understanding the AI/ML Threat Landscape

Join our panel discussion,where experts delve into the unique vulnerabilities of AI systems, including data poisoning, prompt injection, and model extraction. Discover effective strategies for mitigating risks and ensuring robust security in an ever-evolving threat landscape. Gain insights into the regulatory environment shaping AI practices and learn how to safeguard your organization against emerging threats. Don’t miss this essential conversation!

PANEL:

Talking to the Board About Governance for AI, Cyber and Privacy

Join a candid panel offering practical guidance for briefing boards on AI, cybersecurity and privacy. They discuss reporting, focused education and how to prepare to illuminate risk, vendor choices and business value. Hear real world governance models, strategies for "culture change" and an actionable checklist CEOs and chairs can use to demand clear metrics, independent validation and accountable ownership.

PANEL:

Best Practices for Generative AI Risk Management & Security

The panel explores the multifaceted role of AI in enhancing security operations, addressing challenges in compliance, data privacy and ethical AI usage. Experts discuss AI’s dual role in fortifying defenses and being leveraged by adversaries, emphasizing the need for strategic, high-level discourse to guide effective implementation and prioritize AI security initiatives in the evolving cybersecurity landscape.

PANEL:

Best Practices for Generative AI Risk Management & Security

A review of the AI Risk Management Framework and NIST’s 2024 Generative AI Profile which includes 12 unique or exacerbated risks and prescribe tailored mitigations. Hear actionable advice on cybersecurity, third‑party risk, measurement, and governance so your organization can safely adopt GenAI without stifling innovation.

PANEL:

Pushing the Boundaries of Prompt Injection Attacks

In depth coverage of prompt injection attacks and their growing prevalence and complexity. Through examples panelists explore the challenges of defending against these attacks and the need for AI-first approaches and adaptive defenses. The conversation also touches on the limitations of current security measures and the importance of understanding business context to mitigate risks associated with AI systems.

PANEL:

Emerging Protocols: Understanding MCP and A2A Security

Join security leaders and engineers as they unpack MCP’s explosive 2024 adoption, the rise of A2A, and the real‑world risks— prompt‑injection, rogue/local MCP servers, supply‑chain and delegation gaps—threatening enterprise deployments. Panelists share candid deployment lessons, practical defenses like intent‑based permissions, namespace controls and detection‑and‑response, plus a roadmap to get started fast and securely.

PANEL:

Best Practices for Generative AI Risk Management & Security

A review of the AI Risk Management Framework and NIST’s 2024 Generative AI Profile which includes 12 unique or exacerbated risks and prescribe tailored mitigations. Hear actionable advice on cybersecurity, third‑party risk, measurement, and governance so your organization can safely adopt GenAI without stifling innovation.

PANEL:

Securing Your AI Investment

A review of strategies for securing your AI spend without stalling innovation. Our panel shares a practical playbook: quantify risk and ROI, benchmark models and deploy private, zero-retention architectures with AI gateways, guardrails and prompt sanitization. Learn how to align with NIST/ISO and OWASP, leverage CICD and synthetic data and keep humans-in-the-loop, turning domain-specific AI into accurate outcomes while outages and hallucinations are contained—and insurers are satisfied.

PANEL:

Exploring AI Agents: Their Purpose, Capabilities and Impact

The panel discusses how modern AI agents evolve from single task bots to multi-agent digital employees—and what that means for identity, access and risk. Our panel exposes real pitfalls (prompt injection, memory tampering, agent-to-agent drift) and practical controls: zero-trust roles, gateways, AI-over-AI checks and unstructured data governance. Learn playbooks to start narrow, monitor behavior, align with ISO 42001 and prepare for interpretability driven guardrails.

PANEL:

The MAESTRO framework / CSA’s framework

Join Ken Huang and Anton Chuvakin for a discussion on threat modeling Agentic AI using the MAESTRO framework. They unpack how Agentic AI’s non‑determinism and dynamic identities reshape traditional approaches, showcase the seven‑layer MAESTRO framework with a live Gemini‑powered demo and leave you with one actionable tip: start with a clear architecture description to enable automated, human‑guided threat analysis.

PANEL:

Securing LLMs: Practical Protection Strategies for AI Developers

Experts explore the threats and vulnerabilities associated with LLMs and offer practical strategies for AI developers. The conversation highlights the increasing sophistication of attacks, such as data poisoning and prompt injections, and emphasizes the need for robust guardrails and security measures. Panelists also discuss the importance of community collaboration in developing effective risk management practices and red teaming methodologies to safeguard AI systems in various applications.

PANEL:

Balancing Benefits and Risks: The Role of Agentic AI

This panel discussion explores the dual nature of agentic AI, highlighting its potential to revolutionize processes across industries while also bringing a broad spectrum of risks. Experts from AI, cybersecurity, and governance fields discuss the strategic value of agentic AI, particularly in cybersecurity, and emphasize the importance of establishing robust governance frameworks to manage its deployment. The conversation underscores the need to approach the adoption of agentic AI with caution, ensuring thorough risk assessment and stakeholder involvement to harness its benefits responsibly.

PANEL:

Organizational, Operational and Pragmatic Considerations for Securing AI Systems and Data

This panel addresses the complexities of securing AI systems and data. Experts explore the integration of AI into existing security frameworks, emphasizing the need for cross functional collaboration and tailored risk management strategies. The conversation highlights the importance of understanding AI-specific risks, leveraging frameworks effectively and fostering a culture of continuous learning to navigate the evolving landscape of AI security.

PANEL:

Agentic AI: Assessing the Risks and Impacts on Offensive and Defensive Security

Experts delve into the dual role of agentic AI in both offensive and defensive security contexts. They examine the balance between leveraging AI’s autonomous capabilities while implementing necessary guardrails to mitigate risks. The conversation highlights the evolving landscape of AI security, emphasizing the importance of threat modeling, the development of AI-specific security frameworks, and the need for continuous adaptation to safeguard against potential vulnerabilities and emergent behaviors in agentic systems.

PANEL:

Strategies for Red Teaming Agentic AI Applications

This panel focuses on emerging strategies for red teaming agentic AI applications, emphasizing the unique vulnerabilities and security implications associated with these systems. The panel explores the distinctions between agentic AI and traditional LLMs, the necessity of early and scalable red teaming and the importance of processes and metrics for measuring success in identifying and mitigating risks---and the critical need for a collaborative approach in addressing AI security.

PANEL:

Threat Modeling for Agentic AI Applications

The panel discusses the complexities of threat modeling for agentic AI applications and the unique security challenges posed by these systems compared to traditional AI models. They dig into the need for new frameworks to address the dynamic and autonomous nature of agentic AI, exploring potential vulnerabilities, data security concerns, and the implications of agent-to-agent interactions in future AI ecosystems. The conversation highlights the importance of collaboration between security professionals and business units to effectively navigate the evolving landscape of AI security.

PANEL:

The Most Pressing Issues in AI Security: the Current Snapshot

Anton Chuvakin, Ian Swanson, Richard Stiennon and Ken Huang discuss the most pressing threats facing security pros on the AI front—at this moment.

PANEL:

Managing AI Security Risk

This blue chip panel reviews top line strategies for managing AI security risks including continuous assessment of potential vulnerabilities and threats to AI systems, monitoring various attack vectors and evaluating the security of data used to train AI models, the integrity of the models themselves, and the security of the systems where they are deployed.

PANEL:

Winning the AI Arms Race in Cybersecurity

The panel discussion explores the multifaceted role of AI in enhancing security operations, addressing challenges in compliance, data privacy and ethical AI usage. Experts discuss AI’s dual role in fortifying defenses and being leveraged by adversaries, emphasizing the need for strategic, high-level discourse to guide effective implementation and prioritize AI security initiatives in the evolving cybersecurity landscape.

PANEL:

Safeguarding Data Privacy in Agentic AI Systems

Agentic AI supercharges capability and risk. Our panel unpacks real threats (prompt injection, deepfakes, insecure dev tools) and proven privacy playbooks: zero-retention architectures, AI gateways, RBAC/MFA, continuous logging and privacy-enhancing tech (federated learning, homomorphic encryption, differential privacy, zero-knowledge proofs). Learn how to align security, privacy and compliance with ISO 42001/NIST AI RMF, manage deletion requests, govern multi-agent workflows and protect sensitive data at scale.