Ethical Use of AI

Large Language Models and Generative AI are transformative to how we work and learn.

There are a lot of options available when it comes to using AI, including generating CAD models and designs, generating images, data, and code.

Generating Prompts

If you use AI to generate code, you must give it critically engineered prompts that specify what you need from it. As a baseline, you should provide it pseudocode to explain the logic you want from your system. This ensures that the AI is supporting your thinking, rather than replacing it.

When working with hardware, your prompts must also include explicit technical details. For example, if you are connecting an RGB LED, you should specify the signal pins being used, as well as any expected behaviour.

You must keep a record of your prompts, justification as to why you have used those prompts, as well as a critical appraisal the output from the AI.

To appraise the output, consider:

• Did the generated code work first time?
• What problems did you have to overcome implementing it?
• Did you need to manually edit anything in order to effectively use the code?