JAN 2021

Moving towards Human-centered AI

Introduction to the role of Human-centered Design (HCD) in the development of AI-driven services


The design field has spread across several domains and Artificial Intelligence (AI) & Machine Learning (ML) will not be an exception.

This series of articles will be a practical approach to move away from the philosophical perspectives of designing for AI towards a more Human-centered Design (HCD) approach.

I will discuss what contributions, the 'traditional' design process can bring to the design of intelligent systems and how can we move the development of AI services to be more Human-centered. We will see an evolution in the HCD field by moving into coexistence with the formal method of putting human beings in the center of the design process as well as putting data in the center of the design process.
Human-Centered Design, Human-Centered AI, HCD, HCAI, Design Thinking, Design Process, Artificial Intelligence


To achieve social responsibility, designers will have to learn how to design AI-driven services that address issues of fairness, accountability, interpretability, and transparency [2].
Technology plays a significant role in shaping the future of design and is changing the way we think about it. The synergy between design and fields like AI and ML will be essential to assure the success of the established design and AI knowledge. Both fields have the power to shape each other. These developments will require the development of multidisciplinary practices among designers, data scientists, and engineers.
double diamond
Business, design, and engineering integration in the double diamond
So, how can the 'traditional' design process take into account these issues?

To be able to understand how the design process will and is evolving to support AI and ML development, first, we need to reflect on what is Human-centered Design?

Human-centered Design (HCD)

Human-centeredness is a core quality of design. Design has gained special momentum, in the last couple of decades, mainly because of its efficiency towards innovation challenges.

Where did it all start?

First, let's keep this analysis as short as possible. The goal is to reach AI not make a design assessment on HCD history.

During the 1960s, due to technological developments, designers had to shift their focus from Form, as a product of a mere design task, towards the quest of Human Needs as an innovative 'material' to support the ramification of mass production. Design has changed its scope well beyond the conception of artifacts which required a new perspective on the design methods of the time. From this moment, designers are also concerned to understand the intrinsic value of products and services.

In 1969, Nobel Prize laureate Herbert Simon published "The Sciences of the Artificial ", an influential paper on design methods that for the first time formally outlined the process underline Design Thinking. Simon's model presented a process of seven stages, where each component stages and activities has been the cornerstone of the design process for decades.
Simon’s 7 stages of the design process
Simon's 7 stages of the design process
This model was substantially influential in shaping some of the most popular Design Thinking process models used today. Like the five-stage Design Thinking model proposed by the Hasso-Plattner Institute of Design at Stanford (d.school).
HCD timeline
Timeline of human-centered design (HCD) concept
Don Norman massively contributed to the design thought transformation in several domains. But in 1986, Don Norman coined the term User-centered Design (UCD) to describe design based on the need of users. He re-formed participatory design into what became known as UCD.
What happened next, in the late 1980s, was a race to discover new methods for improving business, service, and design. Through this journey, HCD began to transform from a method to a mindset in the early 2000s.

Since the beginning of the 1990s, HCD and UCD were often indiscriminately used to translate the integration of end-users in the design process. Later on, HCD start to evolve when the development of design methods start moving from a techno-driven focus towards a human-focus. The aim was now to humanize the design process while empathizing with users and stakeholders.

UCD reached its golden years when IDEO evolved the HCD approach to what is today highly popularized as Design Thinking in the late 1980s. Although a few designers and scholars had set the stage around this certain mode of thought, it took a few decades to be deep root in business development. Between 1980 and 1983, Bryan Lawson, Nigel Cross, and Donald Schön were the first to theorize this approach. Still, as a mindset and methodology, design thinking is relatively young.
IDEO and D.School at Stanford University

David Kelley
is both the founder of the global design and innovation company IDEO and Stanford University's Hasso Plattner Institute of Design, also known as the d.school.

IDEO and D.School were one of the main channels that made it possible for Design Thinking to reach today's popularity as a tool for innovation.

The design process is what puts Design Thinking into action. It is a structured approach to generating and evolving ideas. The Design Thinking process is a highly interactive and incremental process. The D.School has developed the most popularized Design Thinking model. The model has been designed to be visualized as an iterative series of five stages.
d.School Design Thinking
d.School Design Thinking Model
    • Empathize — This is the beginning of the process. The goal is to gain users, context, and business knowledge through observation and engagement. To empathize we need to view users and their behavior in the context of their lives. Learn about the problem. Engage also with stakeholders.
    • Define — Time and space to identify problems and user needs. Analyze collected data and build user profiles on gathered data to produce actionable problem statements. Express the problem to produce materials that will aid the outlining of our innovation efforts.
    • Ideate — Diverge on possible solutions to address the problem. Evaluate, test, and select wireframe concepts for prototype development. Uncover unexpected areas of exploration.
    • Prototype — Intenerate the generation of interactive solutions that later on will be tested by users. A prototype can be anything that the user can interact with. Create design specifications and evolve concept/wireframes into the design solution to test possibilities, and later on, manage the solution building process.
    • Test — Conduct usability testing and surveys to evaluate design with users in the appropriate context of their life. The goal is to refine the solution and learn more about the users. Tests reveal insights that redefine the problem. Share finding with stakeholders to obtain feedback.

    Moving from HCD to Human-centered AI

    We are not inventing the wheel here. We just need to reframe the HCD model, with a new fresh vision that can meet the necessities of both designers and data scientists.

    Human-centered Artificial Intelligence (HCAI) aim is to move the development of AI systems towards a more human self-efficacy approach. Taking into account issues like fairness, accountability, interpretability, and transparency in its development. HCAI differentiates AI by shifting the center from algorithms and AI systems by replacing it with humans at its center.

    Human-centered artificial intelligence
    Schneiderman, B. (2020). Human-centered artificial intelligence: Three fresh ideas.
    Until recently, researchers and developers focused on building AI algorithms and systems, stressing machine autonomy, measuring algorithm performance, and celebrating what AI systems could do. In contrast, HCAI's grounded on design thinking approach puts the human at the center, emphasizing user experience design, measuring human performance, and celebrating the new powers that people have [3].

    I will leave aside human concerns like privacy, security, environmental protection, social justice, and human rights. Those are fights for another battle.

    Through HCAI more humanistic design research is emerging between data scientists and designers'. Future technological concepts will have to promote human dignity, equality, and security. And we cannot achieve an ethical AI if we don't shift the focus from algorithms performance and metrics to human needs.

    So, how can we shift the design thinking design process towards HCAI?

    Follow Part 2 of this series where I uncover how the 'traditional' design process has shifted to take into account the AI development process. And learn the process to be more Human-centered AI.
    1. Verganti, R., Vendraminelli, L. and Iansiti, M. (2020), Innovation and Design in the Age of Artificial Intelligence. J Prod Innov Manag, 37: 212–227. https://doi.org/10.1111/jpim.12523
    2. Cerejo, J. and Barbosa, A. (2012). The Application of Design Thinking Methodology on Research Practices: A Mind-map of Tools and Methods. 62nd Annual Conference of the International Council for Education Media on Design Thinking in Education, Media, and Society
    3. Schneiderman, B. (2020). Human-centered artificial intelligence: Three fresh ideas. AIS Transactions on Human-Computer Interaction, Vol. 12(3), 109–124.
    4. Riedl, M. O. (2019). Human-Centered Artificial Intelligence and Machine Learning. Human Behavior and Emerging Technologies, Volume 1(Issue 1), 1–8.
    Did you like this article?
    Human + Machine
    Book Review
    Designing the Invisible
    Opinion Article
    Made on