Name: A Step-by-Step Guide to Securing Large Language Models (LLMs)
Start: 2024-10-09T15:00:00Z
End: 2024-10-09T15:00:38.000Z
Location: BrightTALK
Rating: 4.5

We discuss how to identify the full range of risks present against your cloud, on-prem and hybrid data, and eliminate the risks that matter the most.

Let's put data security at the center of information security--where it belongs.

As businesses increasingly incorporate Large Language Models (LLMs) like Microsoft 365 Copilot, OpenAI, Bedrock, and Google’s Vertex into their operations, they face new challenges in protecting sensitive and valuable data. While many companies focus on securing critical data, the real challenge extends far beyond that. The true risk lies in discovering and labeling all data—especially within large AI-driven systems where even overlooked data repositories can lead to significant vulnerabilities.

What you’ll learn:
-Why DSPM (Data Security Posture Management) is essential for every AI initiative—ensuring not only innovation but security as AI tools interact with your data.
-Real-world examples of AI inadvertently exposing sensitive information. Imagine an AI chatbot sharing private customer data or a generative AI accidentally leaking proprietary business strategies. We'll walk you through how these risks can be mitigated.
-A framework for understanding AI data governance and ensuring data security within AI pipelines, so sensitive information is protected, whether it’s stored in databases, hidden in cloud environments, or part of machine learning models.
-A business case for tools like Copilot that enhance decision-making through automation and intelligence while keeping your data secure.

Why This Webinar Matters: In the rush to adopt AI technologies, companies often overlook the complexities of their entire data ecosystem. Data Security Posture Management (DSPM) goes beyond traditional security measures by giving you visibility into all your data, whether known or hidden, ensuring long-term security, compliance, and operational resilience.

Register now to discover how DSPM helps your organization build a foundation for AI that’s both innovative and secure.

Training Your LLM Dragons: Why DSPM is Foundational for Every AI Initiative

If you’re familiar with the children’s book 'If You Give a Moose a Muffin,' then you’ll get this: If you set up Snowflake, you'll spark a wave of user demand for data access. And if you get a wave of user demand for access, you’ll need to decide who gets access to what. But if you give the wrong access to valuable or sensitive data, you’re opening the door to serious risks.

Join us for a live panel featuring leaders from Audacy and Sigma Computing as they share their real-world journeys in securing their data lakes. 

Learn how they’ve maximized their Snowflake investment, provided secure user access, and addressed the challenges of data visibility, governance, and risk. Whether you're tackling data security or looking to optimize your Snowflake environment, this webinar will provide actionable insights from the field.

Unlocking the power of Snowflake: how two customers empower users with the right access and secure their data lakes

In today’s rapidly evolving digital landscape, securing sensitive data has become more critical than ever. With the rise in data breaches and increasingly stringent regulatory requirements, organizations are investing heavily in data security. However, managing vast, unstructured data environments while ensuring compliance remains a significant challenge.

Join us for an insightful webinar where we delve into the findings of the latest research by Omdia, which sheds light on the growing importance of Data Security Posture Management (DSPM) as a foundation for modern data strategies.

During this session, we will explore:

-The increasing allocation of data security budgets in response to an escalating threat landscape and regulatory demands.
-Key data security priorities for organizations over the next 12–24 months, including reducing the risk of data breaches and improving overall security posture.
-The rise of DSPM adoption, particularly among larger organizations, and its role in providing essential visibility and security in data-heavy environments.
-Common challenges organizations face when implementing DSPM, including resource constraints, vendor selection, and gaining executive buy-in.

From Siloed to Unified: DSPM's Transformative Role in Data Security

Data is collected and stored for a reason: to help business teams make better decisions. Yet, so often, there is a battle between giving users the data freedom they need and the risks in doing that. This becomes even more challenging as AI becomes another consumer and generator of data.

As the lines between data management and security blur, the responsibilities of Chief Information Security Officers (CISOs) and Chief Data Officers (CDOs) must align.

In this exclusive panel discussion, Normalyze's customers (Beyond Finance, AccessIT Group, and Informatica) will explore why data and security teams must unite to:
• Safeguard critical assets
• Navigate the increasingly intricate compliance landscape
• And address the new challenges posed by AI in data management and security

Tune in to learn real-world perspectives from your peers.

Moderator: Todd Thiemann, Senior Analyst, IAM & Data Security, Enterprise Strategy Group

Panelists:
Toby Foss, Senior Director of Information Security, Informatica
Matt Murphy, Lead DevSecOps Engineer & Cyber Security Architect, Beyond Finance
Brett Price, Lead Cybersecurity Consultant, vCISO, AccessIT Group

Data Access Versus Data Security: The Tug-of-War Between CDOs and CISOs

Abandoned and shadow data are not just buzzwords; they represent a critical challenge in data management and security. Is it possible that you may be storing data you don’t know you have, some of which might include valuable or sensitive information…all of which is completely forgotten by everyone except cyber criminals?

There are valid reasons why this happens:
-Employees create data repositories in Google Drive, SharePoint, OneDrive, Dropbox, that IT is unaware of.
-Data is backed up without a lifecycle plan, so even when the originating data sources are purged, the backups are forgotten.
-Layoffs, terminations, and other employee activities occur, but no one thinks about the data they leave behind.
-Organizations going through mergers and acquisitions consolidate data repositories, sometimes leaving the original data stores in place.

Join us for an insightful webinar where we delve into the challenges and solutions associated with abandoned data stores. What you will learn during this 30-minute live event:
-The risks and costs associated with data you don’t know you have.
-How DSPM can help security and data teams gain detailed insights into who can access this data, the last time it was accessed, how much of it is valuable or sensitive data.
-Best practices for remediation based on compliance requirements, backup options and more.

This overlooked data costs your organization money and increases security risks. Why not clear out that data to optimize your storage management and improve your security posture? Register now!

Unknown data, hidden threats, wasted spend: tackling abandoned and shadow data

Securing Large Language Models (LLMs) is key in today's AI landscape. This talk dives into viewing LLMs as data compressors, understanding challenges with compressed data, and tracing data origins. Join this session to see how to ensure that LLMs are not trained on sensitive or biased data by implementing on-demand scanning, training automation, and a proxy system to withhold sensitive outputs.

A Step-by-Step Guide to Securing Large Language Models (LLMs)

How is your organization addressing cloud data security? 

Data – including critical, sensitive information such as customer details, financial data and intellectual property – is at the heart of every modern organization. The loss of this data, whether accidental or malicious, will significantly impact your business. How is your organization addressing cloud data security? The traditional approach uses separate point solutions for data loss prevention, governance, data protection, encryption, and threat detection. Siloed solutions often struggle with the growing size and complexity of data. 

The modern approach solves these issues with an integrated single platform described in GigaOm Radar for Data Security Platforms (DSPMs). Join GigaOm Analyst Paul Stringfellow and Normalyze CEO and Co-founder Amer Deeba to discuss GigaOm’s latest findings in their radar report for DSPMs. 

You’ll learn: 
- How well specific DSPM solutions are positioned to serve their market segment and the deployment models needed. 
- What table stakes and critical criteria teams should look for when choosing solutions. 
- How Normalyze addresses real-world use cases with real-time scanning, the most accurate data classification, and AI-powered insights

Analyst Insights: GigaOm Radar Report for Data Security Posture Management

90% of the world's data generated in the last two years is dispersed in multiple isolated environments. As data propagates, security and IT teams lose visibility into and control over where their data resides, how sensitive it is, and whether it is at risk.

How can security and IT teams fix these gaps and regain visibility and control of this sprawling data? How can CISOs assure leadership and shareholders that the organization’s financial well-being and reputation are not at risk?

Todd Thiemann, Senior Analyst at Enterprise Strategy Group, moderates a panel with accomplished CISO and Cybersecurity industry veteran Renee Guttmann, Paychex CISO Bradley Schaufenbuel and ChargePoint CISO Teza Mukkavilli who will share best practices and recommendations for security and IT leaders on building their data security strategies and shared security models.  

Attendees from this session will take away tactics for formulating an actionable security strategy in the context of the shared security model, complying with the new SEC rule (for public companies), and identifying (and remediating) current gaps in their company’s data security strategy.   

Participants include:
Renee Guttmann, Accomplished CISO & Cybersecurity Industry Veteran, CXO Forum
Bradley Schaufenbuel, CISO, Paychex
Teza Mukkavilli, CISO, ChargePoint
Todd Thiemann, Senior Analyst, IAM & Data Security, Enterprise Strategy Group

From Chaos to Clarity: CISOs Share Five Must-Dos to Protect Data

DSPM solutions offer unparalleled visibility into cloud environments, enabling the identification and categorization of sensitive data across platforms. These systems use cutting-edge scanning and detection technologies to identify vulnerabilities, misconfigurations, and non-compliant practices that jeopardize data security. Furthermore, DSPM automates the evaluation and enhancement of data security measures. Continuous monitoring and instant alerts enable preemptive action against emerging threats. Adopting AL and ML enhances these solutions, allowing businesses to rapidly identify potential security breaches before they occur.. Implementing DSPM cultivates a security-first mindset within organizations, underscoring the importance of prioritizing data protection. DSPM not only delivers critical insights but also supports the application of stringent security policies, equipping businesses to defend their cloud data from advanced cyber threats effectively.

Benefits of DSPM:
- Peace of Mind: Detect and categorize sensitive data across all cloud and on-premises data environments for full visibility.
- Breach Prevention: Identify data attack paths; map vulnerabilities, dormant or risky access, and unauthorized activity; and track identities across cloud and SaaS applications to prioritize protections.
- Breach Containment: Monitor anomalous activity and take immediate action against emerging threats in response to prioritized real-time alerts.
- Privacy and Compliance: Automate compliance evaluation to control user access and maintain audit readiness.
- Protect What Matters Most: Address threats, rather than the endless cycle of infrastructure security, to protect your most valuable asset: data.

Modern Solutions for Modern Problems: Securing Your Cloud Data with DSPM

The new SEC Final Rule (effective Dec 18, 2023) requires public companies to disclose security posture annually and cyber incidents within four days after determining an incident was material. Why this new rule? What does it mean for public companies? Join Accomplished CISO and cybersecurity industry veteran Mike Convertino and Normalyze CTO & Co-founder Ravi Ithal on this fireside chat to get insights. 

They will delve into: 
- Future of corporate cybersecurity governance, including board composition, board interest in security postures, non-disclosure to shareholders, and role of cybersec vendors in guiding CISOs. 
- The alignment of interests between boards and the SEC, D&O insurance, and the readiness of companies to prepare legally sufficient security posture statements
- Insights on adaptation timeline, liability insurance for security firms, and security posture statements during IPO disclosures.
- Predictions and trends based on their knowledge and experience.

DSPM for SEC Cybersecurity Disclosure Rules

SaaS applications like Google Workspace and Microsoft 365 have become ingrained in the day-to-day operations of many businesses. Yet, the security risks of these applications have not been fully understood or addressed. Integration of Data Security Posture Management (DSPM) with these applications increases visibility into potential risks and helps secure the data in SaaS apps more proactively. 

Join Normalyze Director of Product Management Vamsi Koduru as he dissects the complexities of securing SaaS applications and illustrates how Normalyze seamlessly fits into the SaaS ecosystem, providing robust solutions to common security risks and gaps.

Key takeaways: 
- Agile SaaS Strategies: Discover short-term and long-term strategies that DSPM and Normalyze offer to enhance the agility of your SaaS applications. 
- Significance of DSPM in Compliance: Understand DSPM's significance in systematizing and securing PII, a cornerstone for regulatory compliance and robust security frameworks.
- Real-World Implementation Insights: Delve into case studies showcasing how customers use DSPM to reduce risk and strengthen data protection.
- Navigating Access Management: Explore why SaaS security controls must also encompass the end-users with access to sensitive data.
- Normalyze's Role: See how the Normalyze product roadmap is evolving with the rapid growth of SaaS platforms.

Register >>

DSPM For SaaS

Join Normalyze VP of Product Management Gautam Kanaparthi and Cohesity Sr. Director of Product Management Sheetal Venkatesh as they discuss an urgent need to protect a company's vital asset—data.

The session will spotlight the integrated Cohesity-Normalyze solution that gives enterprise security and IT teams unprecedented visibility into their most sensitive data and ensures it is automatically protected against data breaches and ransomware.

Gautam and Sheetal will do a product demonstration of how security and backup teams can:

- Identify which sensitive and at-risk data stores have not been protected or backed up.
- Prioritize risk remediation and data protection prioritized by the potential cost to the business if such risks are exploited.
- Remediate risks with data protection and recovery capabilities

DSPM for Proactive Data Protection & Recovery

With the widespread adoption of cloud computing, more sensitive data has moved into virtual environments far from the safe harbor of legacy on-premises data security controls. This raises two critical questions: Is this data secure, and is it secure throughout its lifecycle across the enterprise data environment? 

Data Security Posture Management (DSPM) addresses these concerns by actively discovering and analyzing sensitive data in motion, expediting risk mitigation as data traverses cloud platforms.

Join Normalyze Head of Product Gautam Kanaparthi as he demonstrates how DSPM principles enable enterprises to get a complete view of their data. Key takeaways from the demo will include: 
 
- Data Flow Visualization: See data-in-transit mapping, understand how sensitive data is accessed, and spot policy violations with interactive graphs.

- Data Lineage Insights: Gain insight into how a particular piece of sensitive data has proliferated across your data environment and ensure it is secure everywhere.

- Data access Governance: Continuously assess and monitor data access, enabling swift resolution of dormant, over-privileged accounts and enforcing least privilege for sensitive data.

- Anomaly Detection: Uncover suspicious activity, including data breaches and account takeovers, through continuous user behavior analysis without predefined rules.

- Data Detection and Response: Go beyond configuration checks and detect risky activities around critical data assets leading to data exfiltration

DSPM for Data in Motion

Through interoperable cloud services, attackers defeat native security tools and move laterally through the cloud infrastructure.

In this webinar, Abhinav highlights the real cloud data risk chains that lead to account takeover and exposure of sensitive data and shows what can be done to prevent these attacks.

Examples of adversarial tactics include but are not limited to

* Abusing the account organization’s setup to onboard an external account and making it a trusted entity to avoid detection and perform data exfiltration. 
*Attacker abuses the cross-account trust relation to move laterally between different accounts leading to a fan-out effect while remaining completely hidden from defense tools like Guardduty. 
* An identity-based attack that chains multiple IAM permissions to build an ROP-like privilege escalation scenario and compromise the entire cloud infrastructure.

Top 3 Cloud Data Security Adversarial Tactics and How to Fight Them

No code, data models, or training -- Sigma Computing offers its customers an easy and reliable way to analyze billions of rows of data and make informed decisions in seconds. Sigma users love it. 

For Rahul Gupta, Sigma's Head of Security & GRC, reliably protecting customers’ sensitive data in multiple clouds is the goal of its security strategy. Join Rahul as he shares Sigma's approach to managing data security posture, datastore discovery, access governance, and risk detection -- and how Normalyze enables this framework. 

Co-hosted by Normalyze Head of Product Gautam Kanaparthi, this webinar will cover:

 - Why the security of data in multi-cloud environments is a top concern for CISOs  
 - Tangible, results-based advice for security professionals on building a scalable data security posture framework
 - Insights into cloud-security trends and steps for protecting cloud-resident sensitive data

Securing Customers’ Big Data: Sigma Makes It Easy with Normalyze

Driven by the expansion of cloud adoption, enterprise data is now housed across multiple public clouds, PaaS and SaaS. However, a significant portion remains on premises. As a result, security teams struggle to get holistic visibility into where their data resides, whether it is at risk or compliant, and how much it would cost the company to lose it.

DSPM (Data Security Posture Management) addresses this challenge for ‘data everywhere’ and provides unprecedented visibility into data stores and attack paths across IaaS, PaaS, SaaS, and on-prem environments.

Join Normalyze Head of Product Gautam Kanaparthi as he demonstrates how DSPM gives enterprises a complete view of their data: 
- What type of sensitive data exists and where
- Which data stores contain the most critical information
- Who can access it 
- What potential attack paths can lead to a breach

Gautam will demonstrate how Normalyze prioritizes these risks according to the monetary impact on the business.

DSPM in Hybrid Cloud Environments

Mark provides an overview of Netlify and discusses: 
- access control and reducing the attack surface by controlling and monitoring access to data
- the importance of data classification and security clearances in granting access to applications and data
- compliance and cost considerations related to their access control and security measures

Data Security Challenges: Netlify

AI-Enabled Cybersecurity is Overrated: Why Low Tech IT Security Basics Still Matter in 2025

Unified Defense: The Power of Cisco & Splunk in Data Security and Observability

Tackle 3 Common AI Challenges to Build Effective Security Programs

2024 SonicWall Mid-Year Cyber Threat Report

Navigating AI in Cybersecurity: The 2025 Threatscape

The Promises and Perils of Using AI for Cybersecurity

Purple AI GA: Detect Earlier, Respond Faster, & Stay Ahead of Attacks

Fortifying Enterprise Defenses Against AI-Powered Threats

Unreliable AI: Addressing the Cybersecurity Risks of LLM-Based System

The Primacy of Resiliency: Why (Most) Orgs Shouldn’t Focus on Threats

Ransomware Stopped: Defending Against Fog and Other Threat Actors Now and in 2025

The Evolution of Cyber Threats: How Open Source Tools Are Changing the Game for Better and Worse

Victories, Challenges, and Unknowns: Threat Landscape 2025

What Do You Need to Consider for Your Cyber Incident Response Framework?

Fantastic Cyber Beasts and Where to Find Them: From Current Threats to Future Defenses

Navigating API Security Challenges

Cyber Security 2025: Emerging Threats & Strategic Opportunities for MSPs and MSSPs

Eliminating Cybersecurity Threats: Inspection of DevSecOps

The 2025 Threatscape

AI could contribute up to $15.7 trillion to the global economy by 2030, according to PwC. However, your organization needs to figure out how to secure this technology in order to reap its benefits.
 
In this webinar, Ravi Ithal, Co-Founder and CTO of Normalyze will provide a step-by-step guide to securing Large Language Models (LLMs). 

Key topics covered:
A brief introduction to LLM apps
How LLM security is different and what makes LLMs hard to secure
A framework for securing LLMs
How to implement LLM security measures

LLMs

GenAI

Generative AI

Data Security

Data Protection

Cyber Security

AI Models

Cloud Security

Data Privacy

The IT security community on BrightTALK is composed of more than 200,000 IT security professionals trading relevant information on software assurance, network security and mobile security. Join the conversation by watching on-demand and live information security webinars and asking questions of experts and industry leaders.

IT Security

As an IT professional, many of the problems you face are multifaceted, complex and don’t lend themselves to simple solutions. The information technology community features useful and free information technology resources. Join to browse thousands of videos and webinars on ITIL best practices, IT security strategy and more presented by leading CTOs, CIOs and other technology experts.

A Step-by-Step Guide to Securing Large Language Models (LLMs)

Presented by

About this talk

More from this channel