Chapter 9. Securing AI Services
In earlier chapters, you learned how to build GenAI services that serve various AI generators while supporting concurrency and data streaming in real time. Additionally, you integrated external systems like databases and implemented your own authentication and authorization mechanisms. Finally, you wrote a test suite to verify the functionality and performance of your entire system.
In this chapter, you’ll learn how to implement usage moderation and abuse-protection mechanisms to secure your GenAI services.
Usage Moderation and Abuse Protection
When deploying your GenAI services, you’ll need to consider how your services will be misused and abused by malicious users. This is essential to protect user safety and your own reputation. You won’t know how the users will use your system, so you need to assume the worst and implement guardrails to protect against any misuse or abuse.
According to a recent study on nefarious applications of GenAI, your services may potentially be used with malicious intents, as described in Table 9-1.
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access