The history of user interfaces (UI) in computing has evolved dramatically over the last 60 years, shaping how humans interact with technology. With the advent of generative AI (GenAI), we are now witnessing the emergence of a new major UI paradigm which is redefining how we comprehend and interact with technology. The emerging Artificial Intelligence Experience (AIX) framework aims to integrate these advanced technologies into everyday interactions, ensuring they are intuitive and user-centric.

The Evolution of User Interfaces: From Batch to Intent

The first UI paradigm, batch processing, goes back as far as 1890. In this early model, users submitted a complete set of instructions to computers, initially by punch-card and eventually by program, and awaited the output with no real-time interaction. This mode of operation was rigid and often unforgiving, requiring exact precision in the input to achieve the desired output. Any deviation or error in the inputs could render the entire batch flawed.

As technology advanced, a second paradigm emerged during the 1960’s in the form of command-based interaction. This model introduced a conversational dynamic where users and computers took turns executing commands. Command-based interaction has dominated technological design for the last 60 years due to its increased flexibility. This paradigm facilitated the development of command lines, text-based terminals, and eventually graphical user interfaces (GUIs), significantly shaping the entire field of user experience (UX).

Today, a third UI paradigm is emerging, powered by GenAI technologies like OpenAI’s GPT and Google’s Gemini. This new model shifts the focus from detailed instructions to desired outcomes. Users express their intent, and the AI deduces the best methods to achieve these outcomes, often with minimal input from the user. This transformation not only simplifies interactions but also challenges the traditional role of the user as the primary controller of the process, marking a significant departure from earlier paradigms. We call this paradigm intent-based interaction.

The Challenges & Opportunities of Intent-Based Interaction

The AI user experience represents a different paradigm of interaction between humans and computers — a paradigm that holds much promise, but also highlights significant usability problems.

Unlike its predecessors, with intent-based interaction, users are not required to understand or manipulate the underlying processes of the tasks they wish to accomplish. Instead, they simply state their objectives, and the system interprets and executes the necessary actions. This not only challenges traditional interaction models but also fundamentally alters the locus of control within the interaction, placing the majority of it with the technology.

Consumer GenAI technologies such as ChatGPT and Midjourney exemplify this shift. These systems are designed to generate content, solve problems, and create images based on simple user prompts, bypassing the need for complex command sequences. For instance, rather than a series of commands to create a graphic, a user might simply describe the desired image, and the AI model would generate it according to the specifications. This significantly lowers the technical threshold for users, democratizing access to sophisticated digital tools.

However, intent-based interaction also introduces new challenges, particularly in usability and error handling. The emerging technique of “prompt engineering” has become a necessary intermediary to refine and optimize AI responses based on user input, reminiscent of the early days of search engine optimization where specialists were needed to navigate complex databases. This suggests that while the technology has advanced, there remains a gap between current AI capabilities and the ideal seamless, intuitive user experience.

Moreover, the current implementation of intent-based interfaces often requires users to articulate their needs clearly and precisely in textual form. This presents significant obstacles for those with lower literacy, less familiarity with the technology, or an inability to express what they want — referred to as ‘articulation barriers.’ The risk of misinterpretation by the AI can lead to results that do not meet user expectations, and without transparent processes or explainable outcomes, it can be challenging for users to understand and correct errors. This challenge is further exacerbated by delayed and opaque feedback loops between action (input) and outcome (outputs), which make it difficult to learn from mistakes or improve a user’s experience.

Despite these challenges, the potential for further development in AI interfaces is vast. Future systems are likely to merge the intuitiveness of intent-based interactions with the clarity and familiarity of graphical interfaces, creating hybrid models that could offer the best of both worlds. Such systems will need to carefully balance efficiency and flexibility with transparency and control, ensuring that users can both trust and effectively interact with the technology.

Enhancing AI’s ability to understand and adapt to varied human inputs will be central to achieving a truly intuitive and user-friendly experience.

But as we increasingly design, implement, and interact with AI, our current frameworks for understanding technology interactions prove inadequate. The complexities and nuances of AI require us to rethink our approaches to UX, to ensure they meet the evolving demands of human-AI interaction.

The Importance of Artificial Intelligence Experience (AIX) within AI Interactions

The transition to AI-powered user experiences signifies a fundamental shift in our interaction with technology, comparable to the original computerization movement. This evolution demands a reevaluation of the frameworks we use to understand and design these interactions, ushering in the need for a new paradigm, known as Artificial Intelligence Experience (AIX). Unlike traditional UX, which focuses on command-based usability and momentary interactions, AIX adapts to the agentic nature of AI systems that not only perform tasks but also exhibit a level of autonomy, learning, and engagement in complex social and emotional dynamics.

Artificial intelligence is reshaping how we conceive of technology: not as mere tools serving explicit commands, but as agentic partners in our daily lives. This shift from command-based to intention-based interaction underscores the necessity for AIX. In this new paradigm, interactions with AI are not transient; they are part of an ongoing relationship, which influences and is influenced by the AI’s capacity to understand and adapt to human intentions.

Furthermore, the agency of AI introduces a unique set of challenges and opportunities. AI systems, with their ability to act independently, demand a design philosophy that goes beyond functionality and efficiency. They require an approach that also considers the emotional and social implications of their actions. By engaging in interactions that can mimic human cognitive and emotional behaviors, AI systems challenge our traditional understanding of interaction as a mere exchange of commands and responses. Instead, they invite an interaction style that is more akin to coexisting with another human being, involving negotiations of power, emotional exchanges, and mutual adaptation.

The concept of coexistence with AI brings to light the extended nature of AI interactions, which transcend the immediate and momentary. AIX, therefore, expands the scope of user experience design to include these broader, more sustained aspects of human-AI relationships. It acknowledges that AI technologies are not just tools but participants in these relationships, capable of learning from interactions and influencing them through their actions and decisions.

By embracing AIX, we commit to a design framework that respects and enhances the unique qualities of AI as agentic systems. It is a commitment to prioritizing humans within human-AI interaction, in anticipation of a new era of coexistence with intelligent systems. This requires developing technologies that not only serve us but also enrich our lives by fostering positive, empathetic, and effective interactions. As we continue to integrate AI more deeply into our personal and professional lives, AIX provides the necessary guidance to ensure these technologies contribute positively to our society, enhancing our capabilities without compromising our dignity or autonomy. In this way, AIX does not merely adjust to the presence of AI; it anticipates and shapes a future where humans and intelligent systems collaborate more closely and more meaningfully than ever before.

To learn more about our innovative approach to human-centered AI design and AIX, see our recent case study.

The future of interaction: Generative AI and the evolution of UX was originally published in Bootcamp on Medium, where people are continuing the conversation by highlighting and responding to this story.

In the era of rapid technological advancements, artificial intelligence (AI) has emerged as a transformative force across industries. From frontier models like GPT-4 and Gemini Ultra to applications like Cursor and Shortwave, the list of AI applications is almost endless. In fact, one AI aggregator has over 12,000 unique AI apps at the time of writing.

The quality and accessibility of powerful generative AI (genAI) models has not only increased the opportunity for product innovation, it’s also made product development easier, by providing entrepreneurs, product teams, and developers with powerful tools to support the design and development process. While the focus of many new AI products is on showcasing technological advancements, the success of these applications heavily depends on their user experience (UX).

Doing this successfully requires integrating human-centered design thinking into AI development and implementation, to create products that put the human first. I believe that the future of AI is human-centered, and that these technologies should be designed to enhance and support human interactions rather than replace them.

I call this Artificial Intelligence Experience (AIX).

Within this article I explore AIX via the principles of human-centered design thinking, examining how AIX differs from traditional UX. To explore this new field I draw upon ideas from human-computer interaction (HCI) and UX research, as well as human-centered design principles, founded upon cognitive psychology & anthropological studies of technology. I explain why a nuanced understanding of AIX is vital for successful AI development and implementation.

Human-Centered Design Thinking

At its core, human-centered design thinking prioritizes the needs, experiences, and contexts of people, ensuring that the solutions developed are not only technologically sound but also deeply resonant and meaningful to those who use them. This philosophy is rooted in empathy, a commitment to understanding individuals’ real-world experiences, and a cyclical process of ideation, prototyping, and testing.

The principles of human-centered design thinking are anchored in the belief that effective solutions emerge from a deep understanding of the user’s perspective.¹ These principles include empathy, co-creation, iterative development, and a holistic understanding of the problem space:

• Empathy drives designers to seek out and understand the user’s experiences and emotions, laying a foundation for more relevant and impactful solutions.
• Co-creation involves stakeholders and users in the design process, ensuring the final product is aligned with their needs and expectations.
• Iterative development advocates for rapid prototyping and testing, allowing for continuous feedback and refinement.
• Lastly, a holistic approach encourages looking beyond the immediate problem to understand broader contexts and systems.

In the context of the user experience, human-centered design thinking offers a robust framework for creating systems, technologies, and interfaces that are intuitive and enjoyable to use. This involves employing qualitative and quantitative research methods to gather insights into user behaviors, needs, and motivations. Techniques such as user interviews, surveys, observation, and usability testing become crucial in uncovering the nuances of user experiences. These insights inform the design process, guiding the development of user personas, journey maps, and wireframes that reflect the user’s reality.

Since human-centered design thinking emphasizes a user-centric approach, I also include accessibility and inclusivity, ensuring that products and services cater to a diverse range of users. This approach not only enhances the user experience for a broader audience but also demonstrates a commitment to social responsibility and equity.

In the context of AI, the crucial difference is that systems are able to adapt, respond, and change based upon user needs. Unlike traditional digital products which act as simple tools, products while leverage AI offer a more sophisticated user experience. Rather than a unidirectional command — e.g., a button-press — genAI allows for bidirectional interaction, whether that’s via a text-based chat interface, a voice-based conversational experience, or any other modality. What might seem like a subtle difference creates infinite new possibilities for functionality, and requires a reframing of how we think about the user experience.

Whether determining the tone of voice used by a chatbot, or how a generative feature will be built into an existing product, human-centered design thinking within the context of AI means developing smart systems that are intuitive, adaptable, and responsive to human needs.² It is important to recognize the importance of designing AI that is aware of the situatedness of human actions, necessitating AI systems to be highly adaptable and contextually aware. While traditional UX is deeply rooted in creating interfaces that are intuitive, efficient, and visually appealing,³ AIX introduces a shift towards developing systems that are not only user-centric but also inherently intelligent and anticipatory.

AIX extends beyond the surface level of interaction to foster a deeper synergy between AI systems and users, aiming to create experiences that are truly adaptive and personalized. This evolution from traditional UX to AIX is underscored by the cooperative relationship between human and technology, requiring a more nuanced understanding of human-AI interaction. Unlike conventional UX, where the focus is primarily on the design of static interfaces, AIX emphasizes the dynamic nature of AI interactions, where humans work in conjunction with systems which are capable of learning from and evolve with their users over time. This approach demands a more complex consideration of ethics, empathy, and empowerment in design, ensuring that AI not only complements but also enhances human capabilities without compromising autonomy or agency. In essence, AIX represents a confluence of human-centered design thinking and AI technology, crafting experiences that genuinely understand and adapt to user needs, contexts, and behaviors in real-time.

Enhancing User Experience with AI

Human-centered design thinking becomes increasingly important in the development of AI applications. AI is a powerful technology, which comes in a myriad of forms and flavours. It has the ability to speed up and automate processes, detect anomalies, and generate content. But it also risks being perceived as a threat; removing autonomy and control from the user. If designed well — by putting the human first — AI applications have the potential to revolutionize the user experience by offering more personalized and intuitive interactions. For instance, AI-powered recommendation systems can provide personalized content suggestions, while natural language processing can enable natural and conversational user interfaces. These capabilities, however, must be leveraged with a clear focus on enhancing human capabilities and enriching the human experience, rather than merely showcasing technological advancements. Questions product teams might ask themselves include:

• How might we build our product to maintain user autonomy and agency?
• How might we encourage a meaningful, collaborative relationship between the user and technology?
• How might we embody the principles of human-centered design within our product?
• Does our product play a role within an existing workflow & value chain or require an entirely new one?

I take the perspective that AI applications should support human action and decision-making, and leveraged to enhance cognitive abilities and learning, making interactions with AI not just functional but also beneficial to their users.⁴ This perspective is grounded in real-world implementations of AI. Take, for example, Notion’s implementation of AI. The team behind the release initial assumed people would use their generative AI features to do the work for them, but quickly realized that users didn’t want to outsource the full process to AI. In their own words:

“So we pivoted from AI as content generator to AI as collaborative partner. Rather than replacing your efforts, AI could be a smart guide, helping you get more stuff done, faster and better.”

Focusing on the relational aspect of user interaction with technology — collaborations between humans and AI — leads to the creation of products that users can form a meaningful relationship with, rather than merely use as tools. This relational-focus improves the usability, engagement, and long-term loyalty towards products. But this principle of using AI to support human action and decision-making goes beyond the development of digital products. Through our own work we’ve seen how important this relational aspect is when traditional organizations implement AI solutions. When implemented poorly, AI is perceived as a threat, undermining authority, jeopardizing workflows and harming existing relationships.

Due to this relational nature of AI, the concept of AIX expands far beyond a user interface. The integration of AI into user experience design requires a careful balance between automation and human judgment, adhering to principles which promote human dignity, autonomy, and fairness.⁵ These considerations are vital in ensuring that AI systems respect and promote the human condition, enhancing rather than diminishing a users sense of autonomy.⁶

Case Study: Enhancing Generative AI User Experience

A company building a generative AI creative platform approached me for help in refining its creative suite of tools and UI. The challenge was to understand the evolving needs of a diverse user base. Through a combination of data analytics and user research, we uncovered how users were seeking more intuitive interfaces that could seamlessly integrate into their workflows, while not undermining their autonomy.

The key insight was that users desired “smart” products that could adjust to their skill levels and creative goals. In response to our work, the company developed multiple UIs, tailored to different user segments, enhancing usability and engagement. Additionally, customizable AI response features were introduced, allowing users to adjust outputs to better align with their objectives, and leading to increased user satisfaction and long-term product loyalty.

Conclusion

The integration of human-centered design thinking into AI development is crucial for creating products that truly enhance the human experience. By emphasizing user needs, ethical considerations, and maintaining a human-in-the-loop approach, AI products can become more intuitive, adaptable, and supportive of human endeavors.

Product designers, developers, founders and CTOs should be commitment to this philosophy. It’s paving the way for a future where AI and humans interact in mutually beneficial ways, guided by the interdisciplinary insights and ethical considerations that are essential for the development of a meaningful and empowering AIX.

To learn more about our innovative approach to human-centered AI design and AIX, see our recent case study. To explore what we do, see our services.

Further reading:

¹ Tim Brown, “Change by Design”
² Lucy Suchman, “Plans and Situated Actions.” & Martin Heidegger, “The Question Concerning Technology.”
³ David Benyon, “Designing Interactive Systems: A Comprehensive Guide to HCI, UX & Interaction Design.”
⁴ Benedict Carey, “How We Learn.”
⁵ Sherry Turkle, “The Second Self: Computers and the Human Spirit.”
⁶ Paul Formosa, “Robot Autonomy vs. Human Autonomy: Social Robots, Artificial Intelligence (AI), and the Nature of Autonomy.”