The Values of Value in AI Adoption: Rethinking Efficiency in UX Designers' Workplaces

Through design workshops with 15 UX designers, this paper argues that AI adoption is not merely an efficiency-driven process but a complex negotiation of values across individual, team, and organizational scales that reconfigures roles, power dynamics, and worker agency.

The Gist

Here is an explanation of the paper "The Values of Value in AI Adoption," broken down into simple language with creative analogies.

The Big Idea: It's Not Just About Speed

Imagine a factory where the boss says, "We are getting a new super-fast robot arm! It will make us twice as fast!" The workers are excited but also nervous. They wonder: Will this robot do my job? Will I still need to know how to use the tools? Who decides how we use it?

This paper is about a group of 15 UX designers (people who design how apps and websites look and feel) who went through this exact experience with Artificial Intelligence (AI). The researchers found that while companies talk about AI as a way to be efficient (faster and cheaper), the workers see it as a much more complicated mix of trust, identity, and fairness.

The main takeaway? Adopting AI isn't just a technical upgrade; it's a social negotiation. It changes who does what, who gets credit, and how people relate to each other.

---

The Three Levels of the Problem

The researchers looked at the problem through three different "lenses" or levels, like zooming in and out on a camera.

1. The Individual Level: The "Muscle Atrophy" Fear

The Analogy: Imagine a professional athlete who starts using a motorized wheelchair to get to practice faster.

• The Good: They get to the gym faster and have more energy for the actual game.
• The Bad: They start worrying, "If I stop using the wheelchair, will my legs be too weak to run? Am I losing my skills?"

What the Designers Said:
Designers love AI for the speed. It helps them write drafts, resize images, or organize data quickly. But they feel a deep anxiety.

• The "Prosthetic" Dilemma: They see AI as a tool to help them, but they fear becoming too dependent on it.
• The "Replaceable" Fear: One designer said, "If I automate my whole workflow, will my boss think I'm easily replaceable?"
• The Hidden Labor: It's not just "click and done." They have to spend hours checking the AI's work, fixing its mistakes, and prompting it correctly. It's like having a very fast intern who is also very clumsy; you save time on the typing, but you spend double the time fixing the typos.

2. The Team Level: The "Magic 8-Ball" Problem

The Analogy: Imagine a group of friends cooking a big dinner together. Suddenly, one friend brings in a robot chef that can chop vegetables instantly.

• The Tension: The friends who chop by hand feel their skills are being devalued. The friend with the robot might accidentally serve a dish with plastic in it because they didn't check the robot.
• The Trust Issue: If the robot chef makes a mistake, who is responsible? The person who pressed the button? Or the robot?

What the Designers Said:

• Uneven Adoption: AI spreads like gossip, not like a training manual. Some people are "power users," while others are skeptical or scared. This creates friction.
• The "Easy Answer" Trap: Sometimes, teams stop arguing and debating ideas (which is usually good for creativity) because they just accept whatever the AI generates. It's like taking a shortcut that skips the "deep thinking" part of the recipe.
• Role Confusion: AI blurs the lines. A designer might start doing a developer's job, or a writer's job. This makes everyone wonder, "What is my job anymore? Is my colleague's job safe?"
• Transparency: There's a growing need to say, "I used AI for this part." Without that, trust breaks down. It's like not telling your teammates you used a cheat code in a video game.

3. The Organizational Level: The "Bureaucratic Tug-of-War"

The Analogy: Imagine a school principal who wants to buy a new, high-tech playground. The teachers (the workers) want to use it to help kids learn, but the school board (management) is worried about safety laws, insurance, and budget.

• The Result: The playground sits in a box for months while the principal argues with the insurance company. Meanwhile, the teachers sneak out to play on it anyway because they know it's fun.

What the Designers Said:

• Top-Down vs. Bottom-Up: Management often pushes AI to save money or look "innovative." But the people actually using it (the designers) often have no say in which tools are bought.
• The Compliance Trap: In big companies (like banks or hospitals), legal rules (like privacy laws) move so slowly that by the time the AI is approved, the project is already over.
• The "Box-Checking" Culture: Sometimes companies adopt AI just to say they have it, even if it doesn't actually help the workers. It's like buying a fancy new car just to park it in the garage to impress neighbors, even if it's too expensive to drive.

---

The Core Conflict: Two Types of "Value"

The paper argues that we are using the word "Value" in two different ways, and they are fighting each other:

1. Economic Value (The Boss's View): This is about money and speed. "How many tasks can we finish in an hour?" "How much money can we save?"
2. Social Value (The Worker's View): This is about meaning and relationships. "Does this help me grow as a professional?" "Does this make my team trust each other more?" "Does this respect my expertise?"

The Metaphor:
Think of a Garden.

• Management sees the garden as a Crop Yield. They want to know: "How many tomatoes did we harvest per hour?"
• The Designers see the garden as a Living Ecosystem. They care about: "Is the soil healthy? Are the bees happy? Are we learning how to grow better plants for next year?"

When you only focus on the "Crop Yield" (Efficiency), you might use a chemical that kills the bees (Trust/Collaboration) to get more tomatoes faster. The paper says we need to stop looking only at the tomatoes and start looking at the whole garden.

The Conclusion: What Should We Do?

The researchers aren't saying "Ban AI." They are saying: Stop pretending AI is just a tool.

• It's a Relationship: Adopting AI changes how we treat each other. It changes who is in charge and who is responsible.
• We Need Agency: Workers need a real voice in deciding if and how to use AI, not just being told to use it.
• Rethink "Efficiency": True efficiency isn't just doing things fast; it's doing things well without breaking the team's trust or the workers' confidence.

In short: If we want AI to be a good thing for the workplace, we need to stop asking "How fast is it?" and start asking "Who does this help, who does it hurt, and what kind of team do we want to be?"

Technical Summary

Here is a detailed technical summary of the paper "The Values of Value in AI Adoption: Rethinking Efficiency in UX Designers' Workplaces" by Inha Cha, Catherine Wieczorek, and Richmond Y. Wong.

1. Problem Statement

While organizations increasingly frame Artificial Intelligence (AI) adoption as a critical pathway to competitiveness and innovation, there is a fundamental disconnect between organizational perspectives on productivity/efficiency and workers' perspectives on economic and social value.

• The Gap: Traditional HCI research often focuses on task-level effectiveness or usability. However, the long-term viability of AI in the workplace depends on complex social and organizational dynamics that are often overlooked.
• The Conflict: Organizational definitions of "value" often prioritize quantifiable metrics (speed, cost reduction), whereas UX practitioners view value through the lens of skill development, professional autonomy, rigor, and relational trust.
• The Core Question: How do UX professionals experience, deliberate, and negotiate AI adoption in their everyday work practices, and how do differing notions of "value" surface across individual, team, and organizational scales?

2. Methodology

The study employed a mixed-method qualitative approach involving 15 UX professionals (designers, researchers, and consultants) from various sectors (finance, healthcare, IT, consulting) in the US, Canada, and South Korea.

• Data Collection:
  1. Virtual Design Workshops (5 groups): 2.5-hour sessions using Miro for collaborative activities.
     • Activity 1 (Scenario Building): Participants constructed hypothetical adoption scenarios (e.g., top-down mandates, compliance issues) using a five-stage framework (Discuss, Explore, Test, Select, Adapt) adapted from Rogers' Diffusion of Innovations.
     • Activity 2 (Surfacing Questions): Participants identified critical questions and challenges emerging from their scenarios to discuss collective responsibilities and timing for deliberation.
  2. Follow-up Individual Interviews: 30-minute semi-structured interviews with 14 participants to explore personal reflections, emotional responses, and organizational pressures not fully captured in group settings.
• Data Analysis:
  • Reflexive Thematic Analysis: The researchers used an inductive, interpretivist stance. They analyzed transcripts, Miro board artifacts, and chat logs through multiple rounds of coding, focusing on how participants' lived experiences and positionalities shaped their views on AI.
  • Theoretical Lens: The analysis drew on David Graeber's distinction between value (economic worth) and values (moral/social commitments) to interpret the findings.

3. Key Contributions

The paper shifts the discourse from "does AI work?" to "what does AI mean?" by framing adoption as a process of negotiating values rather than a discrete technical decision.

• Reconceptualizing Efficiency: The authors argue that efficiency is not a neutral technical metric but a situated, contested concept entangled with social dimensions like trust, responsibility, and autonomy.
• Multi-Scalar Framework: The study provides a structured analysis of AI adoption across three distinct but overlapping scales:
  1. Individual: Negotiating self-worth vs. productivity.
  2. Team: Reconfiguring roles, responsibilities, and collaboration norms.
  3. Organizational: Navigating compliance, power dynamics, and cultural norms.
• Critique of "Adoption-Centered" Research: The paper extends HCI's adoption-centered perspective by highlighting how adoption processes often reinforce managerial control and obscure the "hidden labor" required to make AI work.

4. Key Results & Findings

A. Individual Level: Efficiency vs. Self-Worth

• Motivation: Designers primarily adopted AI for efficiency (automating repetitive tasks, accelerating documentation).
• The Paradox: While AI promised productivity, it created anxiety about professional obsolescence. Designers feared that automating their workflow would make them "easily replaceable" or cause their core skills to atrophy.
• Hidden Labor: Participants reported that AI adoption involved significant "hidden labor" (prompt engineering, troubleshooting hallucinations, verification) that often negated the promised time savings.
• Evaluation: Formal metrics (KPIs) were rarely used; instead, designers relied on qualitative, informal judgments of value.

B. Team Level: Reconfiguring Relationships and Roles

• Informal Diffusion: AI tools spread via word-of-mouth and peer advocacy rather than structured training, leading to uneven adoption and "power users" influencing team norms.
• Role Blurring & Anxiety: AI blurred boundaries between UX, content design, and engineering. This created tension regarding job security (e.g., fear that LLMs would replace content writers or junior developers) and emotional burdens for those whose roles overlapped with AI capabilities.
• Erosion of Rigor: A significant concern was the "magic eight-ball thinking" effect, where teams accepted AI outputs uncritically, bypassing necessary debate, conflict, and deep discussion.
• Transparency: There was a strong call for new norms of transparency (e.g., "citing sources" for AI use) to maintain trust and accountability within teams.

C. Organizational Level: Culture and Power

• Top-Down Control: Adoption was largely driven by management goals (cost savings, innovation signaling) rather than practitioner needs. Designers often lacked agency in tool selection.
• Compliance vs. Speed: In regulated industries (healthcare, finance), strict compliance and privacy requirements created a "gray zone" where tools were used unofficially because formal approval processes were too slow for project timelines.
• Cultural Mirror: AI adoption acted as a mirror of organizational culture. In some cases, it reinforced existing hierarchies; in others, it exposed tensions between leadership's desire for efficiency and workers' need for autonomy and safety.

5. Significance and Future Directions

• Theoretical Impact: The paper challenges the HCI community to move beyond "user-centered" design toward a values-centered approach. It posits that AI adoption is a political process where power, labor, and identity are renegotiated.
• Practical Implications:
  • Organizations must recognize that "efficiency" is a contested value, not a universal good.
  • Adoption strategies should include mechanisms for worker agency, allowing teams to deliberate on if, when, and how to integrate AI.
  • New metrics are needed that account for hidden labor, relational trust, and professional development, not just output speed.
• Future Research: The authors call for research into:
  • How AI redistributes responsibility and accountability among team members.
  • Mechanisms for strengthening worker agency (e.g., union-led governance, participatory design) to resist deskilling and ensure AI serves social values.
  • The long-term effects of AI on the "relational labor" of care, trust, and collaboration in the workplace.

Conclusion: The paper concludes that AI adoption is not merely a technical integration but a sociotechnical reconfiguration. To succeed, organizations must treat adoption as an ongoing negotiation of values, ensuring that the pursuit of efficiency does not come at the cost of worker autonomy, professional identity, and the quality of collaborative relationships.