Back to articles

AI-by-Design: A Human-Centred Approach to Building AI Solutions

•2 min read•By Ioannis Zempekakis
AI by DesignFrameworkHuman-Centred AIEthicsPart 1

AI-by-Design: A Human-Centred Approach to Building AI Solutions

This is Part 1 of a 5-part series on AI-by-Design.

We define the human-centred approach to AI — or "AI-by-Design" — as solving the right customer problem in an unbiased, ethical and resilient way (solve the problem right).

The Problem

It is becoming increasingly evident that neither technology nor design alone is sufficient to build AI solutions that work, solve real users' problems, and have a positive impact on society.

  • Data scientists working without designers may move to the solution space too fast, based on unvalidated assumptions, which could lead to solving the wrong problem.
  • Designers with a lacking technical understanding may have an unrealistic view of what AI can do — including that it is somehow "magic", or that it is too difficult to implement.

At the same time, given the dynamic nature of data and AI, it is important to anticipate constant change and design for it.

Our Belief

We believe that AI engineers and designers should collaborate to create solutions that are human-centred, ethical, and have a positive impact.

In this series, we share our most relevant insights from working on AI-based projects and introduce a 6-step AI-by-Design framework to help teams embrace a human-centred approach to AI.

The framework merges the designer's approach to problem-solving (Design Thinking) with the data scientist's expertise in AI innovation. It combines the Double Diamond design approach with the CRISP-DM data management methodology.

The 6 Steps

  1. Discover — Understand customer needs
  2. Define — Find the right problem to solve
  3. AI-by-Design Decision — Should we use AI?
  4. Develop — Explore solutions and design for trust
  5. Test — Validate before building
  6. Deliver & Evaluate — Capture feedback, iterate

Who Is This For?

Our objective is to inspire and motivate managers, product owners, data scientists and designers to start experimenting with AI-by-Design.

Remember: data scientists and designers may have more in common than one might suspect.

Next: Part 2 — Why Do We Need AI-by-Design?