Guidelines for Human Audio AI System Interaction

I designed a set of Responsible Audio AI principles for a AI systems that feed on user data to customize and deliver solutions using audio.

Background

While exploring tentative implementation of Generative AI technologies across the Audio technology space, I along with my teams realized that current AI guidelines do not address the challenges and concerns surrounding audio AI services.

As part of our final deliverable, I took the initiative to craft a set of Human Audio AI Guidelines that future designers and developers can refer to while designing audio AI solutions.

How to Use These Cards

These guidelines are built to assist you across the design process of Audio-AI solutions.

These cards can be used during and after applying RAI principles to your solution. It is important to ensure you are building responsible AI solutions with the help of RAI principles. If you are working specifically with audio, these guidelines are an added check to make sure that your Audio-AI solutions are responsible. You can go through these in any order.

Each card contains a guideline and an example illustrating how the guideline is used in practice. All guidelines may not apply to a single solution. In some cases, some additional guidelines may also be needed, like solutions built with conversational AI. You may even need to make trade-offs between guidelines.

There is no right or wrong way to engage with these guidelines as long as you consider these in your process and work towards responsible Audio-AI solutions.