AI hype vs. AI value
The emergence of AI and its technologies promises efficiency, innovation, optimisation and scalability. For businesses, these are compelling metrics. AI can automate tasks, predict behaviours, and provide insights at a scale that was unthinkable a decade ago. It’s tempting to measure success in these terms, such as how fast the system responds, how much time it saves, or how many processes it automates. But the real currency of AI is trust, and this attribute is intrinsically tied to the human being, the most complex system ever created.
The Paper Straw paradox
But when we design AI to satisfy business goals first, human needs can slip through the cracks, and that trust never gets established, or even worse, trust is lost.
Consider something as mundane as the paper straw. At first glance, it’s a success: it fits the sustainability agenda, fulfils the business requirement, and signals social responsibility. Yet if you’ve ever tried to drink an iced latte or a milkshake through one, it bends, collapses, and frustrates. It meets the goal but fails the user.
This paradox isn’t just about straws. It’s a lens for looking at the evolving world of AI in products. As we rush to adopt AI, we risk creating the digital equivalent of soggy cardboard: systems that are technically impressive, meet business targets, but ultimately leave users frustrated, confused, disempowered and lacking trust.
Human Values as the True Compass
So how do we avoid the paper straw trap in AI? In my own work, I return again and again to a small set of human values that act as practical guardrails, not abstract ideals.
It starts with ease of use. AI should simplify people’s lives, not introduce a new layer of complexity. That means beginning with human needs rather than business requirements. Before designing any AI capability, I ask what experience I want someone to have-how I want it to feel to use-because if the benefit only appears after you understand the system, the design has already failed.
Closely tied to this is consistency and predictability. Trust is built when people can form a mental model of how a system behaves. Even when AI adapts or learns, users should feel they understand what’s happening and why. Designing for consistency reduces cognitive load and anxiety, and it prevents the system from feeling arbitrary or unreliable.
Autonomy is non-negotiable. People need to feel in control, not like passengers in a system that makes decisions on their behalf. That means offering clear choices, meaningful consent, and the ability to override AI decisions when needed. Control isn’t a nice-to-have feature; it’s a prerequisite for trust.
Then there’s transparency. AI doesn’t need to expose its full technical complexity, but it does need to explain itself in human terms. Clear, plain-language explanations help people understand why a decision was made, what data was used, and what options they have. Hidden logic may be efficient, but it quietly erodes trust.
Empathy plays an equally critical role. Human behavior is emotional, contextual, and often messy. That’s why testing with real people matters more than optimizing dashboards. Observing moments of confusion, frustration, or delight reveals far more about the quality of an AI experience than clicks or completion rates ever will.
Finally, there’s trustworthiness, which goes beyond security or compliance. It’s about emotional, psychological, and practical safety. I try to measure not just whether a system is used, but how it affects people’s confidence, sense of agency, and overall wellbeing. Efficiency might drive short-term adoption, but trust determines whether people stay.
Together, these values form a compass. They help ensure that the AI we build doesn’t just work on paper, but works for the humans who live with it every day.
Strategic Responsibility
Everyone working with AI has a strategic responsibility. That should be the conscience, not just the execution. When people interact with interfaces, whether they are physical, digital or AI, the real drivers of behaviour are emotion, effort, habit, and perception.
People make decisions because it feels right, easy, or familiar, not necessarily because the system is technically solid.
Human beings constantly balance effort against reward, comfort against curiosity and trust against risk. When a product, service, or AI system asks for too much cognitive effort or emotional uncertainty, most people simply disengage and stop using it.
That’s why good design isn’t about removing thinking, it’s about removing unnecessary friction. A sense of clarity and control makes people feel safe enough to explore. Predictability and empathy makes them confident enough to trust. When systems behave in ways users can’t anticipate or explain, the trust is broken, even if the outcome is technically correct.
Human decisions are also shaped by habit and perception.
People rely on mental shortcuts, especially in moments of overload. With constant notifications, endless choice, and limited attention, AI must work with these human limitations, not against them. If a system demands too much effort, feels inconsistent, or disturbs the sense of control, it fails. It doesn’t fail because it’s incapable, but because it’s indifferent to how humans actually behave.
At the end of the day, no one enjoys a milkshake through a straw that bends and collapses, or drinking coffee with a faint paper flavour. It frustrates, annoys, and leaves a lasting impression of a system that didn’t consider the human experience. AI is no different. When we ignore how people feel, the effort required, and the small frictions that make or break a moment, we create digital soggy straws. These are solutions that may meet business goals on paper but fail the humans who rely on them.
The cost of the soggy straw
The cost of the soggy paper straw is easy to dismiss because it feels small. It’s just an inconvenience, just a moment of frustration, just a drink that’s slightly less enjoyable. But those moments accumulate. They shape how we feel about a brand, a system, and whether we trust the decisions behind it. More importantly, they reveal a deeper issue: when solutions are designed to satisfy abstract goals rather than lived human experience, they quietly erode confidence. People adapt by working around the problem, lowering expectations, or disengaging altogether. What looks like compliance on the surface often masks frustration underneath.
With AI, the stakes are far higher than an unpleasant coffee. A misaligned AI system doesn’t just frustrate; it can mislead, manipulate, or remove a sense of agency altogether. When efficiency is prioritized over clarity, or automation over consent, people may follow recommendations they don’t fully understand or abandon systems they no longer trust. Over time, this leads to skeptical adoption, reputational damage, and real harm, especially for those already vulnerable.
Designing AI with empathy, clarity, and respect ensures that, unlike the straw, the experience is seamless, satisfying, and trustworthy. It becomes something people actually enjoy using, again and again.
Designing AI for human behaviour means accepting that we’re all flawed. People are imperfect, emotional, distracted, unpredictable, and adaptive. The best AI systems respect this instead of trying to fix us. They reduce cognitive strain, align with habits, and create a sense of safety. Because when design feels effortless and aligned with our instincts, the trust follows naturally.
In a world already crowded and mentally overwhelming, designing for ease, empathy, and transparency isn’t just good practice. It’s an act of respect and love.
About the author
Data Innovation Summit 2026
Raquel Castro is a Portuguese designer with over 17 years of experience, currently based in Norway, where she has worked for the past decade. Her background spans web design, interaction design, and user experience, with a growing focus on the intersection of AI and human experience.
She has designed complex systems including ERPs, finance tools, and B2B software, always prioritizing human-centered thinking over pure efficiency. Raquel’s work is driven by a people-first mindset, bridging technology and empathy to create meaningful attentive experiences that genuinely serve human needs.
At the Data Innovation Summit in Stockholm this May, Raquel will dive deeper into the transition from User Experience (UX) to Human Experience (HX), sharing strategies to design AI systems that value emotional wellbeing as much as efficiency. Join her to discover how to move beyond usability and start building AI-driven products that truly honor the human experience.
*The views and opinions expressed by the author do not necessarily state or reflect the views or positions of Hyperight.com or any entities they represent.
