Beyond the Hype: Getting Real with Human-Centered AI

The Reality

The dream of fully autonomous systems is alluring, but in reality, people crave control and agency. AI should empower users, not replace them. This means designing experiences that preserve human autonomy while providing meaningful interactions. Our user research findings indicate that people prefer a collaborative approach in which humans and machines share autonomy.

The Pivot

Design for shared autonomy. Conduct thorough user research to truly understand users’ expectations and identify opportunities for AI to enhance the user experience rather than diminish human autonomy. Avoid attempting to integrate AI automation fully at every stage of the customer lifecycle. Be more selective, integrating AI where it will add the most value. Offer AI features that enable cooperative user control, preserving a meaningful sense of agency for users. In making this pivot, do the following:

• Start with research. Conduct extensive user research to understand users’ expectations for autonomy to discover real human needs—points at which AI can solve common human problems.
• Map points of shared autonomy. Distill research insights into autonomy maps that balance human, machine, and shared control, identifying points in the user journey that provide opportunities for shared control.

The Method

One helpful method of designing shared autonomy is creating an autonomy map similar to that shown in Figure 1. This form of user journey or service blueprint focuses on mapping the balance of autonomy between users and intelligent systems. Autonomy maps ensure that we maintain human autonomy where it matters most. They also track user and AI sentiment.

Key #2: Relatability

AI relationships, character traits, and actions determine an AI’s relatability for humans.

The Hype

As they attempt to humanize technology, some product makers believe that the more literal the human simulation—whether through avatars, digital twins, or synthetic humans—the more effective the connection that consumers will make with it.

The Reality

People don’t want to be fooled by digital buddies. They want utility, delight, and relatability. Humanizing AI doesn't mean creating lifelike avatars or digital twins. Our research reveals that users often find such literal representations inauthentic and unsettling. People don’t need AIs to pose as humans to form deep, personal relationships with them that they build on real emotions. In short, an AI agent doesn’t need to seem to be human—it just needs to interact in a way that makes sense within the framework of human expectations.

The Pivot

Design for authentic relatability. Embrace relationship design principles and craft AI personas that evoke genuine connections with humans without resorting to artificial human mimicry. Learn from human relationships, but avoid literal representational and emotional anthropomorphism. Recognize that relationships grow over time and must first begin with functional value, then progress later to social value and trust. To pivot from hype to reality, do the following:

• Build trust through utility. We build even the most intimate human relationships on their utility. Earn users’ trust by nailing the basics first.
• Tailor the user interface to deepen bonds. Shape AI interactions and interfaces to the user and the context to deliver personalized experiences that create a real connection.

The Method

A great method that supports designing for relatability is creating affective spectra, which helps identify fundamental AI relationships, attitudes, and personality traits. We often explore different character traits and qualities as the first step in designing an agent, considering various dimensions such as perspective, personalization, familiarity, representation, and emotion, as shown in Figure 2. We then test these agents with users to understand how they respond to these different identities and how comfortable they feel with them and assess how their expectations of the AI have shifted.

Key #3: Immersion

AI user interfaces require different levels of immersion across a various sensory channels, from fully physical to fully digital and a variety of levels in between these extremes.

The Hype

Developing some innovative, novel hardware category and interaction paradigm is essential to creating more immersive experiences that blend AI and spatial computing.

The Reality

The hype around novel hardware and spatial computing often overshadows the fact that we can create truly immersive experiences within existing contexts. Human interactions are inherently multisensory, and we should design an AI to engage the full spectrum of human perception. Although we can expect to see more spatial convergence and the infusion of AI into various contexts, AI adoption is likely to start with the objects and within the environments with which we are already familiar.

The Pivot

Design for natural immersion. Explore multimodal interactions that leverage visual, auditory, and tactile interactions to create more natural, engaging AI experiences. Use AI sensory perception to recognize and take proactive actions and adapt user behaviors in dynamic environments. Attempt to add AI enhancements to current devices and services by doing the following:

• Engage multiple senses. Add social, sensory, and sentiment awareness for deeper and more flexible multimodal interactions.
• Naturalize interactions. Enable both explicit and implicit interactions to allow people to effortlessly use familiar patterns and touchpoints.

The Methods

An immersion spectrum explores concepts across a range of common levels of immersion, from fully physical to fully digital and key points in between, as Figure 3 shows. Sensory swim lanes are a great tool for tracking the complexities of multisensory experiences. As shown in Figure 4, these swim lanes describe an experience flow across multiple, parallel sensory channels over time.

Key #4: Impact

The impacts of AI-enabled user journeys depend on both the value they deliver to users and any negative consequences of the AI’s use.

The Hype

The promise of AI has proclaimed that AI will transform industries overnight and new paradigms will instantly revolutionize human’s work, creating new value. Proponents of AI have touted it as having the potential to solve every human problem. Many believe that businesses must sprint to win the race.

The Reality

The hype of rapid business transformation and value creation has led many companies to rush into AI implementation without having a clear strategy. However, realizing AI’s full potential requires having a long-term vision and focusing on creating measurable impact. While large organizations are starting to realize a measurable ROI from AI, many are spending far ahead of their seeing any return. While efficiency is one useful measure of ROI, increases in revenue are harder to pinpoint.

The Pivot

Design solutions for meaningful impact. Prioritize value over novelty and align AI projects to focus on solutions that address fundamental human needs. Avoid simply adopting theoretical AI solutions in favor of real-time validation. Focus on long-term strategy and growth over short-term gains. Build prototypes using the latest models and platforms to test your assumptions and learn as you go. To ensure that solutions deliver maximal positive impacts, do the following:

• Prioritize value over novelty. Clearly define how AI can solve specific human problems and align your design projects with high-priority, strategic business goals.
• Think big, but start small. Focus on high-value moments to prove hypotheses regarding customers’ needs. Start small, prove the value, and build success by adding new functionality over time.

The Method

When defining near- and long-term roadmaps, we use the opportunity / value framework shown in Figure 5 to assess new features and products that utilize AI. Rank possible AI-enabled use cases according to the value they would bring to users, as well as any adverse effects. You can contrast user-initiated interactions with proactive AI and tedious, low-value work with enjoyable, high-value work, then use AI to automate or enhance tasks, make recommendations, or allow the user to take full responsibility.

Reframed AI: From Hype to Humanity

Within the AI hype cycle, where technological novelty can overshadow substance, there lies a greater opportunity for UX designers to help companies rise above the noise. The genesis of true AI innovation begins with people—understanding their objectives, behaviors, and aspirations.

The AI revolution is not just a technology shift; it’s a shift in human behavior and sentiment. By embracing human-centered design principles and focusing on the four key themes that this column presents, companies can move beyond the hype and create AI experiences that deliver lasting value and impact. 

Credits—I want to credit and thank my team at Punchcut, including Nate Cox, Jodi Burke, and Nick Munro for contributing valuable insights and content to this column.