How I Write with AI

Use AI to increase your value, not stifle your intellect

I use AI in my writing all the time. However, I do not let AI write for me. This does not mean that I do not cut and paste information from AI and use it in my articles. It means that every time I do it, I make sure that I attribute it to the large language model that produced it. However, the ideas, structure, and ultimate content are my creation.

The process of converting my thoughts into words and then writing them is beneficial to my understanding of the material. It forces me to rationally organize my thoughts so that others can easily understand them. I have universally found that after I have written on a subject that I know the subject matter better.

This does not mean that AI is not an excellent aid in writing. I use it frequently. For example, I wrote an article on social media surveillance. I had Claude Sonnet summarize Facebook’s algorithm and Facebook data capture strategy. This was very straightforward and did not require me to rewrite it. However, I wanted the reader to know what I write and what I use from AI. Therefore, I attribute it to Claude and then I cut and paste it in bold. If you look at any of my writing, you can at once tell what AI wrote and what I wrote.

Then there is the question of value. When I was a CEO if I asked someone for some information, I expected them to provide it with their analysis. If all they are going to do is write up prompt for a large language model, what do I need them for? I can certainly write the prompt myself instead of giving them the request, receive the answer quicker and send my questions back to the LLM. Your ability to use your experience in skills to analyze the information is part of your value. I would not willingly give that up.

The following is an example of how I use AI:

Algorithms need your data. They use your data to predict and manage your behavior. Their predictive analytics use sophisticated statistical programs. I asked Claude Sonnet 4 to research the use of this data by social media companies. These are some of the results:

“Major social media platforms and data brokers systematically categorize users through sophisticated psychological profiling systems that achieve 65-87% accuracy in predicting personality traits, creating a $200+ billion surveillance economy that exploits psychological vulnerabilities for commercial and political manipulation. This industrial-scale psychological profiling operates largely without meaningful user consent or awareness, using established psychological frameworks combined with advanced machine learning to analyze behavioral patterns and create detailed personality profiles for targeted advertising, content curation, and influence campaigns.”

I understood that social media companies are profiling us with our data. I asked Claude Sonnet to confirm my assumption. I felt no need to rewrite the Claude response. However, I wanted to make sure that my readers understood that I did not write the information in bold

Here’s another example from the same article:

Social media companies have designed their applications to capture as much granular data as possible. The application stores your data in a centralized profile. They analyze this data and sell it. The buyers use it predict and manipulate your behavior. Here is an example of the information that Facebook keeps on its users:

This is the response from Claude 4 when I asked it to research what information that Facebook kept.

“Facebook’s algorithms store user data in sophisticated multi-layered systems designed for both comprehensive profiling and efficient retrieval. Here’s how they manage individual user data:

User Profile Data Structures

1. Feature Vectors - Each user is represented by high-dimensional vectors containing thousands of attributes tracking interests, behaviors, and demographics.

2. Graph Databases - Social connections are stored in specialized graph databases that map relationships between users, content, and interactions, optimized for traversing complex connection patterns.

3. Engagement Histories - Chronological logs of interactions (clicks, views, time spent, reactions) are maintained with temporal decay factors to prioritize recent behavior.

4. Topic Affinity Scores - Numerical representations of interest levels across thousands of categorized topics, continuously updated based on engagement.”

Claude gave a very clear definition that was unnecessary to rewrite. Once again, I wanted to make sure that the readers know that this was from Claude.

This process has sped up my ability to author articles and blogs while keeping my value in the process. I retain the benefits of organized, thinking while having the AI benefit of speeding up research, providing factual definitions of the accepted concepts as well as proofreading my final document. AI can be a multiplier of your intellect, or it can stifle your intellect. Do not allow AI to reduce your creativity. Use your creativity to make AI work better for all of us.