Designed Intelligence: Achieving the full potential of human and artificial intelligence
This article was originally published on Design Voices, our design blog. Find it here. By Connor Upton Artificial intelligence (AI) is already firmly enmeshed in our lives and businesses the world over. It sits behind many of our apps and services, making our busy lives manageable through prediction, personalisation, and automation. We once asked how can we apply artificial intelligence, but the question now is how do we use AI to its full potential? It’s a big question — not just for businesses aiming to improve services but for each of us as individuals.
As we grow ever more reliant upon it, there is a growing critique of its pervasiveness in our lives and its unintended consequences. At times, AI systems have failed to make good decisions because they can’t quite fathom the complexity and context of the real-world.
There are also issues around AI dependency when people become over-reliant on technology. The most dramatic examples of this are autonomous vehicle accidents, where drivers don’t have full awareness when the system needs them to step in.
Failures like these deeply affect the lives of individuals, they can result in reputational damage for businesses, and at a societal level they reduce trust in technology and data. They stem from a mindset that perceives AI as automation technology and a business case that seeks to replace human activity to improve efficiency. In reality, most things that can be automated will be automated. After all, we don’t want to be hand weaving our clothes or managing the spam in our inboxes. But as AI becomes broadly adopted and is applied to more complex problems we must be careful to understand the unique value that people bring. People tend to not just define useful processes but are also remarkably flexible, for example developing communications and hacks that keep organisations running. If we don’t learn from these, we run the risk of making systems less resilient and increasingly risky. By the same token, paying close attention to the contribution of the human in AI systems allows us to better confront and account for our own harmful biases.
“When designing artificial intelligence it’s critical that we also design for human intelligence and the interactions between these agents.”
Beyond the risks, taking a more human- and systems-oriented approach to AI brings huge opportunities. What if we looked beyond cost reduction and adopted a growth mindset? Organisations are driven by people and their ingenuity. So, the fundamental question should be: how might we use AI to amplify people’s abilities to create net new value? Achieving this requires a new systematic approach.
We call this approach Designed Intelligence. It encapsulates three aspects of how businesses can harness AI to generate value responsibly. First is the application of AI to envision new strategies, second is harnessing it to empower people in increasingly complex and dynamic systems, and third is using it to enhance the human experience.
The three E’s of Designed Intelligence.
Envisioning new strategies
AI-inspired Tapas from the Dock.
Artificial Intelligence is not like human intelligence. It sees the world in a very different, data-oriented way. This alternative perspective can help people break out of their orthodoxies and envision new strategies, services, and products that they would never be able to imagine on their own. We use this approach in projects and design thinking sessions to help businesses blend human expertise with analytical insight, giving them what we call “informed intuition”. We’ve applied this to help food manufacturers develop new flavour combinations, to help banks imagine new personal services for financial wellness, and to help hotels redesign how they measure quality.
Empowering people within systems
While digital services have made our lives more convenient, they can also make systems difficult to understand and less transparent. For customers, this includes understanding why you’re getting that product recommendation, or not getting that loan approval. For workers, it might be figuring out why your caseload is changing or deciphering your performance report. As AI becomes more prevalent in digital services, their logic could become even less clear as algorithms are typically designed for outcomes, not explanations.
This poses a challenge for businesses that want to apply AI. While AI has the potential to bring great benefits, it will only succeed if people are willing to adopt it. In our research, we find that trust between people and AI tools requires transparency, control, and relevant interactions. To get the best out of an AI investment, we help businesses understand how their people are currently working, what pain points they encounter and how AI will change their practices.
For example, in our work on the Accenture Logistics Platform, we helped postal dispatchers harness AI to predict service demand and optimise delivery routes. We also designed a mobile app to guide delivery workers in the field. Workers can make adjustments to the AI’s suggested routes directly in the application and even override them if needed. This means workers can build up trust in the service over time while remaining empowered. Crucially this ability to act independently is also a source of new training data that allows the AI models to get feedback, adapt to changes and improve.
Interface from the Accenture Logistics Platform. Click to watch video.
As the world evolves, businesses will continue to need people long into the future. People can deal with ill-defined problems, interpret the real world with all its messiness and bring context to complex issues like social biases and discrimination. At the same time, AI can radically scale our human expertise by speeding up mundane tasks, monitoring dynamic systems and discovering relationships in large complex data sets.
Enhancing the human experience
At the experience level, we’re applying AI to extend human capabilities. Computer vision and voice technologies are forms of AI that provide more natural ways to interact with information. Applied correctly they can create automagical experiences that truly engage customers. Working with VELUX, for example, we built a mobile app that lets people visualise what new roof windows would look like in their home. This intuitive service gives homeowners the ability to envisage 3D space like an architect and provides them with the confidence to make dramatic changes to their homes.
Designed Intelligence is about helping businesses embrace the next stage of AI innovation where people and machines generate value together. This approach makes the invisible visible and enables people to experience what AI can do for them and with them. As Kevin Kelly say in his book, The Inevitable, “this is not a race against the machines, this is a race with the machines”.
Designing for both human and artificial intelligence together will build trust, increase adoption and open up a myriad of ways to solve challenges on a scale and complexity like never before.
“It’s not a race against the machines, this is a race with the machines.” — Kevin Kelly, The Inevitable