Business efficiency and human collaboration
Empowering the creative economy through design
AI has the power to reshape our world in unimaginable ways. But will this transformation elevate or eclipse human experience? The answer lies in our commitment to a human-centric approach that prioritizes empathy, ethics, and equity.
Most would agree that AI must enhance, rather than complicate human lives. A human-centered approach embraces this perspective from the outset. It focuses on identifying and solving problems that directly impact each one of us individuals. It ensures that technology serves a purpose other than showcasing technological capabilities. Applying human-centered AI (HCAI) to every enterprise problem fosters innovation by prioritizing the needs and values of individuals, ultimately benefiting the organization, its people, and the broader society. Let’s remember that.
Every AI problem is a human problem. It’s not just a technology problem.
From input-output to intent-outcome: The paradigm shift with AI computing
With AI, technology has transitioned from essential input-output functions to sophisticated intent-outcome interactions. What does this mean to us?
Take, for example, a simple app that we use to set up an alarm clock every morning. Before AI, we gave exact instructions to the computer asking it to wake us up at a specified time. AI can now predict user needs based on statements like “I need to wake up early,” eliminating the need for instructions to set an alarm. AI can make recommendations for us with more insights (from our data) than we possibly know ourselves. Imagine an AI recommending improving sleep by modifying room lighting, suggesting dietary changes, or asking you to switch off your devices 30 minutes before bedtime.
AI the healthy living guide
The shift from ‘input/output’ to ‘intent/outcome’ requires us to deeply understand human intentions to ensure that the AI we design is purposeful and serves the needs of people. This means that the less obvious parts of AI, which are not immediately visible to users, are just as significant and influential as the more visible and familiar aspects of traditional user interfaces. People often ask what is there to design in AI. We mistake the design for the cosmetic chat or the recommendation interface when the role of design goes deeper than the surface. Design for AI is about decoding the invisible needs of people, their aspirations, their beliefs, and their values with which they interact with technology.
Designing for AI is about decoding how humans work and encoding it into how AI works.
Characteristics of human centered AI
1. Focus on supporting human decision-making
Consider this: An individual makes approximately 35,000 decisions daily – 30 million annually. In a 1,000-employee organization, this number amounts to 13 billion decisions each year. While not every decision would benefit from AI, optimizing decision-making intelligently can save time for more creative work for people. AI empowers better insights based on data, allowing us to make quality decisions. When the role of AI is to enhance enterprise decision-making for people, there is no question about AI taking over humans. AI is in service of humans.
Real-world applications of HCAI have markedly improved decision-making and business outcomes. In the consumer-packaged goods industry, AI plays a pivotal role in decision-making, using extensive data analysis to refine pricing and promotional strategies, thereby boosting profitability and responsiveness. In the healthcare sector, HCAI is helping streamline insurance claim reviews by pre-evaluating medical records for missing information or suggesting more economical treatments, giving healthcare professionals the leeway to focus on informed decision-making. In manufacturing, computer vision technology on production lines monitors processes, cutting down the need for manual checks and allowing staff to address root quality issues directly.
2. Putting Human in control of the AI automation ecosystem.
When we talk about responsible AI (RAI), we often focus on fixing data biases. We forget that truly Responsible AI is about taking responsibility for the entire experience a human has with AI. The role of design is to ensure that humans are not just in the loop, but in control of the outcomes given by AI.
AI can operate at different levels of automation, from non to fully autonomous systems. For example, in the retail industry, a pricing analyst may decide all prices at the lowest level (no automation), or let AI manage pricing entirely at the highest level with no human involved in choosing, approving, or overseeing. However, we do know that setting a price for a product is an informed, albeit sometimes irrational, decision based on several factors such as the experience and expertise of the human, context of operations, processes, policies, governance, and even cultural norms, brand identity and so on. It is so important to design the optimal of automation and optimal human control, so the risks of AI automation are reduced significantly (think about the spectrum of damage caused by AI from loss of jobs to loss of lives).
A simple design exercise: See the 10 levels of automation (Source: Human-Centered AI, Ben Schneidermann) below and identify the right level of automation that would benefit your use case without overdoing automation + giving the right level of human control.
3. Strengthen human mastery and human connection
Technology has always been a bicycle to our mind to improve our creativity and productivity. This demonstration of seeking mastery can be seen across every industry that applies AI today in manufacturing, CPG, retail, telecommunications, healthcare, or media. The goal is for AI to empower us, enhance our abilities in our jobs, and make us masters in our job.
Consider our AI product, Qure.ai, tailored to aid radiologists. It utilizes deep learning for improved imaging accuracy and automated interpretation of the diagnosis of X-rays to find fractures or signs of lung cancer. It enables radiologists to spend more time caring for patients choosing the right treatment for them, as opposed to spending hours in front of a scanning machine.
Another critical yet often overlooked aspect of AI is our innate need for social connection. Technology has been a huge enabler of bringing people together from the invention of the internet to social media today. Bringing people together will become an essential role for AI applications both in the consumer and in the enterprise world. In fact, that will be the key differentiator between what is perceived as good AI and bad AI.
We took Qure.ai to the Everest Base Camp (an 8-minute video here), a location that has no medical facilities for the 50,000 people trekking or living in the mountains. Equipped with Qure.ai, a single radiologist and an AI specialist could swiftly diagnose and treat severe lung conditions due to high-altitude living. Through AI we connect to people we could not connect to earlier and provide quality healthcare.
Qure.ai diagnostic solution
If AI is not connecting to people and meeting their needs, it is simply a novelty and will fade away as a fad. There are many GenAI tools today that had an initial surge in adoption and are now losing interest with people, simply because they are unable to serve the needs of people and be purposeful.
The purpose of AI is in its ability to serve people’s unmet needs.
As we bring AI into different areas, putting people first is essential. At Fractal, we take a human-centered approach to solving any problem with AI with a team of neuroscientists, psychologists, designers, strategists, engineers, and domain experts. At the core of everything we do at Fractal, is our need to solve human problems, not just data problems. By maintaining a human-centered approach and balancing automation with human oversight, AI can enrich our professional and personal lives and be an empowering ally to us.