Responsible AI Use: Maintaining Human Control in Creative Work

At JetStyle, AI is a regular assistant in our workflow. It helps us research, edit, and explore ideas. We carefully test all new frameworks and keep only the most useful ones for ourselves and our clients. At the same time, we need a code of conduct to ensure that we remain fully human in our decision-making and communication with colleagues and partners.

Our CEO and co-founder, Alexey Kulakov, one of the ambassadors for AI integration at JetStyle, put together a set of principles to maintain clarity, ownership, and responsibility when using AI.

1. Make all conclusions yourself, and be ready to retrace the path to the conclusion

It’s the most important rule: every conclusion must be made independently, with your own mind. No conclusion should be made by the machine for you. Be ready to explain aloud how you reached the conclusions you present to colleagues. You need to go through the path to the conclusions yourself, and own those conclusions.

2. Share only what you wrote 

AI can assist extensively, but anything you share with your colleagues should be written or edited by you. Giving your team access to raw AI documents is fine, but what you intend others to read should reflect your own reasoning. Include references if needed: “These are the sources I used to reach this conclusion; they are available if you want to check.” 

3. Explicitly mark AI-generated draft sections where you have doubts

If you reference a document written by an LLM, treat it as a draft. First, read it yourself and explicitly mark all sections where you have doubts, highlighting what seems important. Do this before passing the information to others. Otherwise, it is unclear whether you have actually thought through the text and whether it can be trusted.

Example: 

Revenue will grow 40% next quarter. // unsure: figure from AI response, original source not verified

This way, the reader immediately sees which parts you vouch for and which parts are uncertain.

4. Do not present AI work as your own

Mark clearly what you did yourself and what was done by artificial intelligence. Acknowledging AI contributions is OK; there is no shame. But lying about it destroys trust.

5. Do not cite AI as an authority

Under no circumstances should you claim “artificial intelligence said this” as reasoning that supports your position. We use borrowed knowledge from the machine, but referencing LLM outputs as an argument is professionally incompetent. “AI said so” is the best way to cast doubt on your own competence.

6. Store context in Git and make it usable

Store your project context in Git: sources, intermediate materials, AI outputs, decisions. To make the context genuinely usable, you need minimum organization effort, such as: search, a convenient interface to the knowledge base, and a bot to help interact with the team’s knowledge base.

7. Keep your work aligned with the goal at all times

At every moment, you should maintain the connection between what you are doing with AI and your core goal. Keeping your activity aligned with the goal is the main thing to care about, even without AI. When you interact with a machine, this becomes especially important, because AI often tempts you to expand the boundaries of the task. So keeping focus gets even harder. Humans are supposed to set the framework: why we are doing this, what exactly needs to be done, which criteria I use to check the quality, and which constraints must not be violated. 

8. Research begins with source selection

Immediately after you formulate a hypothesis and define the goal, task, and criteria, take care to limit the list of sources to work with. A research request without source selection will produce unspecified results.

9. No quality criterion = no value from AI

If you cannot measure the quality of the result yourself — and there is no way to know whether what AI delivered serves your goal — you get arbitrary quality. And if you get arbitrary results several times in a row, you guarantee poor outcomes.

To see these principles in practice, explore our other AI resources at JetStyle, where we share workflows, frameworks, and examples of responsible AI use in creative and XR projects: https://jet.style/articles-ai 

In any project with us — whether it’s design, XR, or product development — we ensure that AI is used effectively to maximize your results, save time, and optimize resources. The core priority is to also keep human judgment central to the process.