My AI Partnership: why and how
About my use of AI:
- ❌ I usually use AI to improve my writing process: to check spelling, accuracy, flow, grammar, and tone.
- 🧐 Even where other sources are quoted, I cross-check with Claude.ai and other secondary sources for validation.
- 💡 My writing process usually involves finding something inspiring, interesting, coincidental, absurd, or odd. Then, I work collaboratively with my preferred AI, Claude.ai. Why? My fingers don’t work like they did. My mind doesn’t put ideas together like they did. Claude helps me care for my readers. No one wants to eat a half-baked pie. Similarly, no one wants to spend time wading through the waves of poorly connected drivel that popilates the internet. I owe it to my readers (both of you!) to form my ideas thoughtfully, to provide context, to add details, and to present other sides. Every single day of writing, Claude helps me do that. I owe it to you to use Claude (or other AIs).
- 🦮 I work with my AI collaboratively. This means that Claude looks over what I’ve written. It highlights the strengths. It also identifies opportunities. This is just like when I’m working with a person I like: I don’t have to hide because the AI isn’t coming from a critical place. I propose ideas and Claude responds with comments, counter-proposals, and other ideas. Usually, what I was first thinking forms the germ of further thought. In most cases, many further thoughts! I’m a little older so sometimes I need a little boost getting up from the floor. My ddog is more than happy to help steady me (for a price!). That’s why I began this point with a dog. I like to think of Claude as another kind of a dog (or intelligence) that wants to help.
- 📚 Claude is not a person. My use of AI builds on my strengths. It enhances my humanity. I cautions me when I might be pushing people away rather than drawing them near. My use of AI preserves, rather than detracts, from the humanity of this si te. By virtue of reading through endless documents, looking for patterns, looking for reception, Claude helps me be a better human being. Even kinder. AIs are another kind of intelligence, an alien intelligence.
- ⚡ Do I think AI is dangerous? Yes. Like a firearm is dangerous. It can provide food, art, beauty, or harm us and those who fall into using it to extract, harm, coerce, and create pain. A choice lies before us.
- 👩🏽🔬 AI is fundamentally a human endeavor. The creators of every AI have chosen to erect guardrails to prevent misuse. The questions about guardrails are: who erects those guardrails, what guardrails exist, why they exist, and how they are modified when misuse occurs.
- 🥽 It matters what AI you use. AIs should be designed to be helpful, honest, and harmless (see the excellent writings of Ethan Mollick). AIs should have anticipatory guardrails. The reason I use Claude.ai is because it’s governed by a Constitution with continually modified anticipatory guardrails (those it can anticipate) with open research on how it makes mistakes.
- 🦺 My use of AI is focused on safety not extraction or exploitation. I chose Claude because it seems the safest AI there is. One day, as all human things, it will chose a different path like many other AIs currently do. I will have to make other decisions at that time.