June 2026
Intermediate
392 pages
11h 24m
English
You might wonder why I’ve relegated the topic of AI safety to Chapter 8. After all, isn’t the point of improving AI’s security to make it safe? In fact, the term “AI safety” refers to a distinct discipline with its own set of concerns. While security typically focuses on protecting a system from external threats such as adversarial attacks, data poisoning, or model theft, safety, by contrast, centers on preventing the system itself from causing harm to its environment, whether through unintended behaviors, misaligned goals, or overreach.
With the rise of LLMs and agentic AI, the line between intentional threats and unintentional ...
Read now
Unlock full access