AI in Code Development: Make It Fast AND Secure

With an uptick in vibe coding and AI-assisted coding practices in general, organizations are finding their teams are trying to take advantage of the speed and efficiency of AI in development, but this is often at the expense of security. LLMs lean towards functionality over security in code generation tasks. Generated code can contain vulnerabilities such as missing input validation, SQL injection, weak authentication, hardcoded credentials and outdated cryptographic algorithm use. These risks are further amplified by unvetted training data, insecure prompting patterns and a lack of standards for integrating SAST/DAST into AI-assisted development pipelines. Add to that the growing threat of prompt injection, data poisoning and model supply chain compromise, and the attack surface expands fast. In this webinar, IANS Faculty Joff Thyer breaks down the areas of risk to focus on and provides recommendations to put more security guardrails back into place.