AI Con USA 2025 - Security
Customize your AI Con USA 2025 experience with sessions covering security for software developers and testers.
Tuesday, June 10
Prompt Engineering for Software Practitioners
With the sudden rise of ChatGPT and large language models (LLMs), practitioners are using these tools for all aspects of engineering. This includes leveraging LLMs for creating software artifacts such as requirements documents, source code, and tests; reviewing them for issues and making corrective suggestions, and analyzing or summarizing results or outcomes. However, if LLM's are not fed good prompts describing the task that the AI is supposed to perform, their responses can be inaccurate and unreliable. Join Tariq King as he teaches you how to craft high-quality AI prompts and...
Thursday, June 12
Securing the Foundations of AI: Addressing the Past to Safeguard the Future
AI’s future hinges on an ecosystem built on decades of technical debt, fragmented tools, and opaque processes; creating vulnerabilities that threaten the reliability and security of modern applications. In this talk, Peter will examine how the legacy of open-source numerical computing and software supply chains is influencing AI’s trajectory. Drawing from over a decade of leadership in the Python and scientific computing communities, Peter will share strategies for tackling these challenges: improving transparency in data and dependencies, building curated software stacks, addressing...
Shadow Vulnerabilities in AI/ML Data Stacks - What You Don’t Know CAN Hurt You
The adoption of open-source AI software introduces a new family of vulnerabilities to organizations. Some components in AI, like model serving, include Remote Code Execution (RCE) by design, like when loading pre-trained models from external sources. Traditional SCA and SAST approaches are not built for the AI ecosystem leaving a huge & insecure attack surface. The irony is that in the AI ecosystem, security issues such as remote code execution are actually a feature and not a bug, often specified explicitly in the docs, which most devs don’t read. AI models are often downloaded from...