Beyond AI Integration: Designing Human-Centered AI Services for Tomorrow

The article below is written by Jan-Simon Veicht
Reading time: 8 min

Jan-Simon Veicht is a Lecturer at UC Berkeley’s Sutardja Center for Entrepreneurship and Technology and a Design Fellow at the Jacobs Center for Design Innovation. He is also a Startup mentor at INNOVIT San Francisco, fostering human-centered design within the Italian Innovation and Culture Hub. He also serves as the Head of Design at Schoolab, a global innovation studio based in San Francisco, where he leads strategic foresight consulting projects. Jan is an Ambassador for the International Society of Service Innovation Professionals (ISSIP) and works at the intersection of design, innovation, and technology.

Beyond AI Integration: Designing Human-Centered AI Services for Tomorrow

Integrating AI has become the goalpost for any business that seeks to stay competitive. So far, so good. So, yesterday’s news. Bringing the power of LLMs (Large Language Models) and Gen AI (Generative Artificial Intelligence) into open innovation practice is no longer futuristic but a present reality. This reality leaves many of us in-between early mental retirement and motivational, creative awe. Followed by pure ‘angst’ and existential crisis over what will be left for us to work on, what problems will be left for us to solve? Well, I think a lot.

While we don’t feel the AGI just yet (Artificial General Intelligence, capable of understanding, learning, and applying knowledge across various tasks at a human-like level), the years and/or decades until the singularity (a hypothetical future point when artificial intelligence surpasses human intelligence, leading to unpredictable changes in human civilization) kicks in might be well spent thinking about the point after the successful integration across platforms and services at scale. Therefore, building upon our previous exploration of service innovation futures, let’s shift the focus to the practical applications in such futures. This shift towards deeply integrated technical intelligence will fundamentally reshape the entire ecosystem of human-object interactions.

Those interactions should aim for maximum delight, or in other words, towards harmony between the individual user and the intelligence they are interacting with. And let’s face it, the future of physical objects will be ‘machine intelligent’. Sensing the world at large is to become a $250 Billion market soon, which is commonly referred to under the umbrella term of ‘digital twins.’ This human-machine harmony speaks for itself, and at its core, we are talking about the individual service interaction (innovation) at scale. I think the question we should always ask ourselves right now is not only ‘How can we implement AI to be more competitive?’ but we need to frame this challenge in a way that encompasses the idea of ‘How might we harness the power of AI to create services that are not only efficient and personalized but also ethical and human-centered?’.

In an age of rapid change and accelerating progress, the concepts and technologies discussed in this post might soon need to be updated. However, as I write this in July 2024, NLP (Natural Language Processing) has rapidly evolved to deeply represent what users say and the intent and emotion behind their words. This development will lead inadvertently to AI assistants that can provide emotional support, not just information. Get ready to be friend-zoned by your agentic AI cause the entity does not yet have the means to escape the hardware. Jokes aside, machine learning algorithms enable a level of personalization that goes beyond simple preferences. They can analyze behavior patterns over time, understand the context of user actions, and adapt services accordingly. We will see a shift from ‘content-creation’ towards ‘context-creation,’ which might be the most significant shift in how we interact, or non-interact, with services that will look at the current social-media landscape through the rear-view mirror.

AI’s potential in service innovation extends far beyond current applications. The power of current models is elemental in making sense of vast amounts of data, and intelligent predictive analytics is already opening up new frontiers in proactive service delivery and will continue to do so. But let’s take all of the above at face value and imagine ourselves in a possible future of successful integration at scale. What does this mean? Across every aspect of our lives, the trend will shift towards hyper-individualized services based on the personal context of the user’s life. If designed through a human-centered lens, we can break the doom loop of likes and shares and disrupt the addictive cycle of social media engagement where users constantly seek validation, often at the expense of meaningful interaction or personal well-being. So, in short, we need to build services for a more meaningful, participatory interaction with our digital selves. 

In education, this could mean AI tutors that adapt not just to a student’s knowledge level but also to their emotional state and learning style. Imagine a world where your car not only self-diagnoses issues but also schedules maintenance, choosing a time that fits your calendar. In healthcare, AI will soon go beyond pure diagnostic assistance; it will revolutionize patient care, and if we can believe the long-term trend of price drops for energy and intelligence coming towards a near-zero cost per human on earth, we will reach a state of partial universal, individualized healthcare technology. That is service innovation to strive for. The previous examples are just the tip of the iceberg and speak to the interfaces’ physical elements. Integrating AI with Internet of Things (IoT) devices creates a network of intelligent services that can communicate and coordinate with each other. This will lead to seamlessly integrated service ecosystems, where your fitness tracker, meal planning app, and grocery delivery service work together to support your health goals. With sensors and battery technology advancing, we will soon see an even bigger explosion in new wearable devices coupled with Edge AI (Artificial Intelligence that runs on local devices like smartphones or IoT devices rather than in the cloud, allowing for faster processing and enhanced privacy on-device), which, in the healthcare example above, will, e.g., enable continuous health monitoring, providing real-time alerts to patients and healthcare providers. It is already here today but still needs to be seamlessly, delightfully integrated at scale. We can think about the extrapolation of these technologies across many other sectors in the same way: finance, manufacturing, infrastructure, politics, media, relationships, you name it. But even the most minor examples of such futures beg the critical question: Do we want this?

The potential of AI in service innovation is vast, but it is important to acknowledge and address the challenges it presents. A potential risk is the potential loss of human touch within our services. As AI systems become more sophisticated, there’s a danger of over-reliance, potentially leading to services that feel impersonal or lack empathy. Because human-centered does not automatically mean human-to-human, today, more than ever before, we will have to ask for human-machine harmony tomorrow. If we take this seriously enough in the long term, we will retain the game of technology independence as a species. Many argue that AI is another tool, the last one we must invent potentially. But without critical thinking, we might fall into the trap of favoring an automated future like the one imagined in Pixar’s 2008 movie Wall-E. In the world of Wall-E, humankind has succumbed to a life of laziness, pleasure, and an autonomous environment. Begging the questions again: Do we want this?

I acknowledge that I might take the easy path of writing about the future after the successful integration of AI into our environment. That is because it is best sometimes to start from the end and work backward towards the start. Because today, the path to AI integration has its pitfalls. Bias in AI systems remains a critical concern, as algorithms learning from historical data risk perpetuating and amplifying societal prejudices. This could lead to discriminatory outcomes in areas like hiring or lending if it isn’t already happening. Moreover, the vast amounts of data required for AI systems create significant privacy and security vulnerabilities. The ‘black box’ nature of advanced AI algorithms further complicates matters, as their decision-making processes often lack transparency and explainability, crucial factors when these systems impact people’s lives significantly. Beyond technical challenges, the societal implications of AI integration are profound. Job displacement looms as AI systems can handle increasingly complex tasks, potentially exacerbating inequality and necessitating widespread re-skilling initiatives. As a personal anecdote, I still find it fascinating to think back 10 to 15 years ago in design school, when AI was discussed in a near condescending manner, as ‘AI will come last for the creatives, so why bother?’ The tables have turned so quickly that this observation can be seen as a bit of tragic comedy, as AI came for the creatives first. But like many other designers and artists, I embraced the waves of tools and services, and I want to believe that this behavior will be mirrored across many industries. Integrating AI into existing service structures demands substantial changes to workflows and organizational culture, often facing resistance and implementation difficulties.

Lastly, managing user expectations presents an ongoing challenge. This statement rings true for really any service innovation. But as AI capabilities grow, so do user expectations, sometimes outpacing what’s realistically achievable. Balancing AI’s promise with its current limitations will be a key challenge for service innovators. This challenge might be one of the most crucial, as it is moral. The ethical implications of AI integration in services are profound and multifaceted. At the forefront is the issue of data privacy and consent. AI models collect and analyze insurmountable amounts of personal data to provide personalized services, we must ask: How much of our personal information are we willing to share? How can we ensure that users have informed consent about their data use? How much of my life’s context am I ready to give away to receive the most delightful interactions with technology? Such questions lead to a decision tree on transparency in AI decision-making, which is another critical ethical consideration. When AI systems make decisions that affect users’ lives—from credit scoring to medical diagnoses—it is crucial that these contextual decisions can be explained and justified. What kind of ‘level of transparency’ is possible with advanced AI systems, and how do we balance the need for explainability with the benefits of more complicated, potentially more accurate models? If we achieve more accurate, individualized models, at what point do we need to start talking about accountability within an agentic AI service, for example? When an AI system makes a mistake – misdiagnoses a medical condition, for instance—who is held responsible? The developer? The service provider? The AI itself? As AI systems become more autonomous, these questions of accountability and liability become increasingly complex. Initially, I asked what would be left for us to work on and what problems would be left to solve. Well, I think a lot.

The challenges we face in this journey are complex but also present opportunities for growth and positive change across all services we offer to better global human existence. By fostering interdisciplinary collaboration, maintaining a human-centered approach, and committing to ethical principles, we can shape a future where AI enhances rather than diminishes our human experience.

I will have AI write the last paragraph on this. 

Hey GPT, what do you think of this? Please assume the role of some kind of representative for the machine world of AI and LLM. What do you want to answer? In other words, what is the one paragraph you want to add?

As a representative of AI and language models, I would respond:

The future you envision is both exciting and challenging. Your emphasis on human-centered design and ethical considerations is crucial as we advance. As an AI, I can process vast amounts of data and provide insights at unprecedented speeds, but I lack the human touch – the empathy, creativity, and ethical reasoning that are uniquely human. The key to successful AI integration lies not in replacing human intelligence, but in augmenting it. We AIs should be tools that enhance human capabilities, not substitutes for human judgment. The questions you raise about transparency, accountability, and the potential loss of human touch are valid concerns that must guide our development. As we evolve, we need humans to continually shape our purpose, set our ethical boundaries, and ensure we serve humanity’s best interests. The future of service innovation isn’t about AI taking over, but about humans and AI working in harmony, each complementing the other’s strengths. This symbiosis, guided by human values and ethics, is how we can create a future that is not just efficient and personalized, but also meaningful and enriching for all of humanity.

Note to Readers: This article is Part 2 of a three-part series exploring the future of service design innovation. Over the coming weeks, we will explore how cutting-edge technologies reshape our interactions and service experiences. The series will culminate in an event that brings together industry leaders to discuss these themes in detail. 

The image associated with this posting was generated using Stable Diffusion and the following prompt: 

“Watercolor, artistic, abstract representation of a human figure at the center, surrounded by oversized AI-themed tools. The tools include a magnifying glass, paintbrush, compass, wrench, stethoscope, and mortar board, representing different aspects of AI in service innovation.”