Olivier Bourgeois

Olivier Bourgeois

Software developer with a focus on cloud, back-end development, and developer experience.

My stance on AI

In an era of commodified artificial intelligence, I want to be fully transparent on how I feel about AI and where AI has been used (or not) within my projects. See something that I forgot, or that you disagree with? I would love to hear your thoughts.

Conversations

In general, I think using AI tools as a substitute for what would otherwise be a search engine query is okay, and I do use Gemini that way. I take the same precautions as for search engine queries (e.g. is the output verifiable by a source of authority?)

Yes: Using AI to brainstorm ideas

Yes: Using AI to ask questions and learn

Yes: Using AI to help link thoughts together

Code generation

I have GitHub Copilot set-up in my coding editor and use it primarily like I would Stack Overflow (e.g. “how do you iterate over a dictionary, again?”) I keep my usage sparse and never use code that I do not understand or that breaks consistency.

Yes: Using AI to code-complete lines.

Yes: Using AI to generate small functions (< 10 lines).

Yes: Using AI to auto-complete boiler-plate or repetitive code.

No: Using AI to generate code I do not understand.

No: Using AI to generate large chunks of code (> 10 lines).

Content creation

I have never published content generated by an AI, and do not intend on ever doing so. Beyond the variety of societal impact it causes[1], misrepresenting myself or claiming AI generated content as mine goes against my values.

There are variations of a quote floating around that says, “why should I bother reading something no one could be bothered to write?” I’d like to propose a further variation to encompass much more than just written content: “why should I bother appreciating something no one could be bothered to create?”

To me, art is deeply personal and is one of the few sacred wonders and joys of the human experience. If we outsource its creation to computers… What do we even have left?

Yes: Using AI to critique work and ask for feedback.

Yes: Using AI to validate consistency of written content.

No: Using AI to generate written content (blog posts, emails, etc.)

No: Using AI to generate visual arts (photos, paintings, drawings, etc.)

Training AI using copyrighted material

Under no circumstances do I condone the training of AI models using copyrighted material without proper authorization from their respective creators. I find it disgraceful, and I oppose using AI models that have been found to engage in doing so.

No: Using AI trained on copyrighted material.


This page was inspired by Damola Morenikeji’s AI manifesto and is listed on SlashAI.

[1] I recommend reading Johan Brandstedt’s “Top ten lies about AI art, debunked”, Erik Hoel’s “Here lies the internet, murdered by generative AI”, and Phil Venables’s “Where the Wild Things Are: Second Order Risks of AI”.