Engaging and empowering end-users has been a fundamental research objective of our Center for Lifelong Learning and Design (L3D) for the last few decades. My contribution will document how a simplistic and narrow initial understanding has evolved based on identifying the critical roles that end-users can and need to play to address the wicked problems that individuals and communities are facing in the digital age. Grounded in the analysis, the findings, and the assessment of specific frameworks and socio-technical developments, research objectives, requirements, and design trade-offs are described to further refine the concept of end-user in the age of AI ensuring that technological innovations contribute positively to cultural evolution in an increasingly digital world.
1.1. Pinball Construction Set: The Importance of Domain Knowledge In the “stone age” of personal computing (before 1.5 million Apps were available on Apple laptops), we experimented with the Pinball Construction Set — a video game by Bill Budge written for the Apple II and later released by Electronic Arts in 1983 (see Figure 1).
Our experiments with experienced programmers and sophisticated Pinball machine players resulted in the finding that with minimal training the Pinball machine players constructed more interesting and sophisticated games than the experienced programmers with limited knowledge about playing with Pinball Machines.
These findings provided the rationale for the design requirement that application-domain
knowledge and support environments for human problem-domain interaction [
The design insights and requirements grounded and derived from the interview shown in Table 1 and other empirical and theoretical work included:
Putting Owners of Problems in Charge: Ill-defined problems cannot be delegated; therefore the owner(s) of a problem need to be present in incrementally framing the problems, because they have the “authority” to change the problem. If owners of problems are in charge, then background assumptions do not need to be fully articulated [29]. It is a strength of domain experts that they know the larger problem context, which enables them to solve ill-defined design problems, to learn while solving problems, to notice similarities between design problems, and to know when design rules can and should be broken.
Support for Unselfconscious Cultures of Design: Alexander [
Domain-oriented design environments (DODEs) (1): DODEs put end-users (being the
owners of problems) in charge by supporting
o human problem-domain interaction rather than just human-computer interaction
[
The Seeding, Evolutionary Growth, Reseeding process (SER) Model (3): its focus is to
incrementally refine and evolve systems as living entities [
Related frameworks to end-user development (complementing the ones mentioned in the previous
section) have been described in [
End-User Programming (EUP) focused on the objective of empowering and supporting
endusers to program (with techniques such as: programming by demonstration, visual
programming, scripting languages, and domain-specific languages)[26];
End-User Software Engineering (EUSE) adding to EUP support for systematic and disciplined
activities for the whole software lifecycle (including: reliability, efficiency, usability, and
version control) [
The book "Democratizing Innovation" [32] contributes to the framing of the concept of end-users by demonstrating with specific examples that end-users are not just the target audience for technological products but are integral to the innovation process itself. The book highlights the importance of open innovation and collaborative design processes, where users and producers cocreate solutions.
In the societal world at large in which governments create rules and regulations, the concept of “Nudges” (based on human behavior and psychology) [31] postulates the principle of libertarian paternalism in which officials create rules and regulations to guide human behavior. In doing so, they act as “choice architects” (in analogy to meta-designers in our frameworks) influencing people's behavior in a predictable manner without restricting their freedom, autonomy, empowerment, and welfare of choice for the citizens (the “end-users” in this framework). In this approach, default settings are chosen not just for convenience but to guide users towards more efficient and beneficial options. Citizens retain their freedom of choice, but the design subtly influences their decisions in a direction deemed positive by the designers.
In the Age of AI (not just in the last few years but more or less from the very beginning) two distinct approaches originated and emerged with separate traditions, values, priorities, and visions in the computing world [17][23][28]: • • one being (strong) Artificial Intelligence (AI) to replace human beings, automating the human experience, and duplicating human behavior with computing systems; the other being Intelligence Augmentation (IA) to expand and complement human abilities with sociotechnical environments.
The emergence of generative AI and large language models (LLMs) during the last 18 months is transforming how the role of end-users is perceived, expanding their capabilities, changing their expectations, and reshaping their interactions with technology.
Research objectives and design trade-offs (focusing on ChatGPT as an example) that need to be
critically examined include promises such as that it will
change our minds about how we work, how we think, and what human creativity really is
[16];
enrich our understanding and increase the support for “distributed cognition” frameworks
[
Equal attention needs to be given to analyzing pitfalls and hindrances [20] such as its fluency is an illusion that stems from the combination of massive amounts of data, immense computing power, and novel processing techniques; its limited support for asking questions and framing problems; the difficulty for humans to scrutinize the answer produced; the fact that it hallucinates and gives wrong answers; its unreflective use leading to humans suffering from an overreliance on technology. • • • • • • • • • or •
The design tradeoffs and balances between these advantages and disadvantages often depend on the contexts in which ChatGPT is used, the specific needs of the users, and the evolving capabilities of ChatGPT. As these tools continue to evolve, addressing the disadvantages while enhancing the advantages represents a fundamental challenge.
A Narrative for Different Uses of ChatGPT: Personal Speechwriters. High-level politicians, industry leaders, and other decision makers are using speechwriters to assist them in giving presentations on a broad range of different topics. If one or more speechwriters provide a text and the “important persons” deliver the speech by reading the text literally given to them — they engage in the least demanding activity: all they need to be able to do is read.
What the listeners of the speech hope for and expect is that the “important persons” superimpose their own ideas based on the ideas provided by the speech writers.
Most people writing an essay, a job application, or an admission letter for a university cannot afford to have another human as a speech writer. They can use ChatGPT as their personal speech writer in two fundamentally different ways:
Behavior-1 (undesirable): use the text created by ChatGPT as their final product (governed by an “AI versus Human” perspective) Behavior-2 (desirable): use ChatGPT to generate a first draft and then rewrite the texts to reflect their own voices and experiences (governed by an “AI and Human” perspective). In this case, ChatGPT can be a powerful tool by blending its computational power with human intuition, expertise, and ethical judgment.
A core objective for the 2024 CoPDA Workshop “Differentiating and Deepening the Concept of “End User” in the Digital Age” should be to explore what different AI approaches can contribute for enriching our understanding of empowering end-users in the future and which developments will be detrimental.
One drawback is that humans may be forced to cope with the burden of being active contributors
in personally irrelevant activities that can lead to participation overload as illustrated by (1)
“do-ityourself” societies (e.g., companies offloading work to customers [
The concept of the "end-user" has transformed from a simple role of consumption or operation to a complex, dynamic participant in the design, development, and use of products and services. This evolution reflects broader changes in technology, education, society, and the global economy, underscoring the importance of continuously adapting our understanding of who end-users are, what they need, and what they can contribute.