Page 14 - CyberSecurityEssentialsEnglish
P. 14

List of Figures
         Chapter One: What Is Artificial Intelligence?
       • Figure 1-1: Acceleration, augmentation, and automation as productive roles
       • Figure 1-2: Approaches to building AI systems and how they relate

       • Figure 1-3: Task families by output type
       • Figure 1-4: From data to training, inference, and updating

       • Figure 1-5: How an answer is generated step by step
       • Figure 1-6: Where each limit bites in a single task
       • Figure 1-7: The jagged technological frontier
       • Figure 1-8: A collaboration loop that keeps decisions and responsibility with people

         Chapter Two: Talking to AI, the Craft of Prompting
       • Figure 2-1: From an idea to a tested prompt
       • Figure 2-2: Ask, instruct, and collaborate

       • Figure 2-3: The anatomy of a professional prompt
       • Figure 2-4: Prompt, source, or tool?
       • Figure 2-5: The PACFT framework

       • Figure 2-6: Diagnose the cause before you rewrite the prompt
       • Figure 2-7: From a text format to structured outputs

       • Figure 2-8: A workflow for long documents
       • Figure 2-9: The lifecycle of a prompt template
         Chapter Three: Inside the AI Market
       • Figure 3-1: A first decision tree for choosing a type of tool

       • Figure 3-2: The layers of an AI system
       • Figure 3-3: The probability distribution and the choice of the next token
       • Figure 3-4: A simplified map of the AI market

       • Figure 3-5: Choosing the intervention according to the cause of failure
       • Figure 3-6: The full path of a trustworthy retrieval system
       • Figure 3-7: The tool evaluation cycle before and after adoption

       • Figure 3-8: Why language and fact must be scored separately
       • Figure 3-9: The seven-factor matching framework, gates then trade-offs

         Chapter Four: Using AI Safely Without Losing Control
       • Figure 4-1: The 4D framework for safe and productive use
       • Figure 4-2: The 4D framework as one cycle of four questions
       • Figure 4-3: The ladder of delegation and human control

       • Figure 4-4: The data minimization path before sending
       • Figure 4-5: Prompt injection and jailbreaking are not the same attack
       • Figure 4-6: A verification pyramid proportional to the impact of error

       • Figure 4-7: The SIFT method for verifying information and sources
       • Figure 4-8: Four measures of fairness, and why they collide
       • Figure 4-9: What to do with a suspect file, in order



         13
   9   10   11   12   13   14   15   16   17   18   19