Be the change you want to see in the artificial intelligence world. Or scramble to catch up.

Hope Is Not Lost for Human-Centered AI
How designers can lead the charge in creating AI that truly benefits humanity.

The rapid proliferation of Artificial Intelligence (AI) brings with it a range of ethical and societal concerns. From inherent biases in datasets to fears of widespread job displacement, these challenges often feel like inevitable trade-offs as AI becomes deeply embedded in our lives. However, hope remains. Human-centered AI—designed to be fair, transparent, and genuinely beneficial—is not only possible but achievable when crafted with intentionality.

For UX professionals, this is an opportunity to drive the creation of AI systems that empower rather than overshadow human capabilities.

---

A Quick Note on AI Literacy

To make meaningful contributions to AI product development, designers need a foundational understanding of how AI works. While a PhD in machine learning isn’t necessary, being an informed practitioner is essential.

Think of learning about AI like learning to invest. At first, it seems daunting—what even is an ETF? But with time, the jargon and processes become familiar. Similarly, while you don’t need to be a machine-learning expert to work with AI, understanding its basics is critical.

AI refers broadly to a computer’s ability to mimic human thought, while machine learning (ML)—a subset of AI—enables systems to learn from data. Unlike traditional programming, where explicit instructions are coded line by line, ML models identify patterns within training datasets. These models then function as “black boxes,” generating outputs based on user inputs—though the inner workings are often opaque.

Understanding these fundamentals empowers designers to bridge the gap between AI’s technical potential and its real-world application.

---

Design-Led AI

Ideally, designers are involved from the very beginning of AI product development—during the discovery phase. Here, we evaluate whether AI is the right solution for a given problem, ensuring user needs drive decisions rather than the allure of flashy tech.

Key questions to ground AI solutions in user needs include:

• Is AI the best tool for solving this problem?
• How will the model integrate into existing user workflows?
• Can the necessary data be provided in real-time?
• How should outputs be presented for optimal usability?

Basic AI literacy allows designers to make informed judgments and collaborate effectively with engineers. Engaging early ensures that AI solutions are designed to adapt to users—not the other way around.

But what happens when design isn’t brought in until after AI decisions have been made?

---

Design-Guarded AI

Even when AI is a foregone conclusion, designers can still shape outcomes by focusing on the two areas where users interact directly with AI: inputs and outputs.

Input Design

Whether inputs involve transaction data, images, or text prompts, the method of collection must be intuitive and user-friendly. Established design principles, such as affordances, help ensure clarity and simplicity.

For example:

• Fraud detection: Automate data collection to minimize manual effort.
• Health screening: Provide clear instructions for capturing images, with real-time feedback to guide users.
• Generative AI: Offer pre-defined options for prompts to reduce the paralysis of a blank input field.

Frequent user testing ensures input methods align with real workflows and pain points. The result? Streamlined, user-centric experiences that reduce friction and save time.

Output Design

Designing outputs requires a focus on transparency and mitigating automation bias—the tendency to over-rely on AI. Users must understand that AI is fallible.

For instance:

• Fraud detection: Avoid overwhelming users with unnecessary alerts to prevent “alert fatigue.”
• Health screening: Display confidence scores (e.g., “80% confident”) to encourage critical evaluation of results.
• Generative AI: Include references or citations to allow users to verify content, presenting outputs as tools for decision-making rather than absolute truths.

AI should act as a collaborator, not an authority. Outputs must empower users to make informed choices while supporting their next steps within a seamless workflow.

---

Ethics Must Take Center Stage

No discussion of human-centered AI is complete without addressing ethics. Designers must champion transparency, inclusivity, and fairness throughout the product lifecycle.

Questions around bias, privacy, and unintended consequences must be raised early and revisited often. While ethical considerations may sometimes conflict with short-term business goals, prioritizing them is essential for building AI that serves humanity in the long term.

These conversations won’t always be easy—but they are necessary. As designers, we have the tools and responsibility to ensure AI remains a force for good. By advocating for human-centered design principles, we can help shape an AI-powered future that enhances human potential rather than undermining it.