My dad's request, albeit with added politeness, echoes decades of entrenched search engine interactions. It underscores how, while I’m soaring high on cloud AI, most are still standing at the base of the mountain.
AI will be transformative for those with knowledge and access. It could potentially give rise to a new kind of disparity, where those with access to AI and the understanding to use it effectively become a privileged class, further widening the gap between them and those who lack such advantages.
As designers, our mission is to create experiences that are accessible and easy for all users. But Large Language Model (LLM)-driven AI presents us with a new challenge. This technology isn't an incremental new feature. It's about shifting the entire paradigm of interaction. We're transitioning from passive tools to thinking machines, systems that don't just respond to input but actively engage with us, seeking to understand and work on our behalf.
This evolution will be profound. It's not something that we can achieve merely by introducing AI capabilities to existing applications and expecting users (like my dad) to figure it out. AI is not an addition; it's a transformation, and this transformation must be user-centered.
How do we make such a significant interaction shift accessible for everyone?
Designing successful digital products involves more than just shaping interfaces—it means crafting interactions that work to educate. These experiences will be fundamental to mass-market AI adoption; how do we foster the process in a natural way?
From my own learning journey, I believe three key approaches will be pivotal:
We design AI to teach its users, in a human-centered way, how to best interact with it.
Unlike its 90’spredecessors, LLM-driven AI can analyze our intent and interactions to provide deeply personalized feedback, suggestions, predictions, behaviors, and more.
In the instance of my dad’s 'salmon recipe, please,' LLM-based tools will come back with, “Here’s a salmon recipe for you, and by the way, you can always try asking for a recipe with specific ingredients or one that caters to dietary restrictions.” It doesn't just fulfill his request; this guidance nudges him toward a more intricate interaction with the AI, deepening his understanding and enhancing his experience.
Another crucial approach is designing with deep empathy for the user's perspective, focusing not only on the technical side of AI but on the user's experience, needs, and context.
Empathetic UX design ensures we're creating for every user's reality, be it the tech-savvy young adult, the curious beginner, or my dad - each interacting with AI for the first time. This strategy considers the diversity of user scenarios, pushing designers to think beyond the confines of tech-centric design and instead cater to a broader audience.
Consider a mobile banking app. To a millennial, the design is intuitive - they can transfer money, check balances, and pay bills with ease. But for my dad, it may seem daunting. An empathetic design approach would prioritize simplifying these tasks, minimizing technical jargon, and creating an intuitive interface that guides first-time users through the process.
For AI novices like my dad, empathetic design elements might not only provide his salmon recipe but might also consider his digital comfort level. It could offer to read out the recipe aloud, convert measurements based on his preference, or even suggest pairing the salmon dish with a suitable wine. This scenario presents a design that considers not just the immediate request but also the unspoken needs and context of the user.
Our final design strategy focuses on controlling the complexity presented to the user at any given time. It starts by showing users the basics. Then, as they become more comfortable, it gradually introduces more complex functions and features. This disclosure of information will look completely different for each individual user based on their past interactions.
Take Duolingo, for example: the language learning app initially presents users with basic words and simple sentences. As they gain proficiency, they're gradually introduced to complex language structures, additional vocabulary, and new features like stories and podcasts. This approach, adaptable to user performance, allows learning at a comfortable pace, ensuring mastery of simpler concepts before moving on to more challenging content.
Imagine a future scenario: My dad, now an AI-savvy user, prepares for a family event. He boldly asks his voice assistant for a flavorful salmon recipe. The GPT-like AI suggests one with a Brazilian twist, and then it nudges a new feature into the mix—an AI-powered cooking app for step-by-step visual instructions. With this function now in his arsenal, down the road it might detect when he’s short of ingredients and suggest an AI-integrated grocery app to update his shopping list. And so his education continues.
For this reality to take shape, humanity needs to not only be comfortable interacting with AI, but willing to learn from it. To grow with it. What does the future look like when everyone can navigate this new AI landscape effectively? And how will our interactions with AI evolve as our understanding deepens? It’s our job to create experiences that constantly evolve to users' expectations.
AI will transform society. That change will be weighted in favor of those who know how to use it and the designers that make it easy to do so. By leveraging evolved design strategies, we're steadily bridging the knowledge gap between novices and enthusiasts, making conversational AI accessible to all—simplifying our collective journey, one salmon recipe at a time.