Artificial intelligence (AI) is arguably the most exhilarating advancement in technology today. The potential of AI to transform every touchpoint between the physical and the digital realms surfaces with growing regularly in the media.
Sometimes the applications are surprisingly delightful, often genuinely useful and occasionally downright scary. From writing catchy pop songs to brewing the perfect beer, from optimizing agricultural yields to personalizing education for school children, from determining sexual orientation through facial analysis to waging war with autonomous weapons, the bandwidth of potential AI applications – and their ethical ramifications – are indeed broad. The idea that a non-biological creation can learn, solve complex goals, and flourish in our world is a mighty leap forward. Where will it take us? How will we steer this emerging powerhouse that so many in the world are creating and evolving together and in parallel?
Allowing people to develop AI without any constraints clearly does not appear to be the best option. For this reason, a number of big tech companies, among them Google, Microsoft, and SAP, have stated their positions on AI and drawn up their own principles for addressing the ethical issues associated with it. As a business leader, a designer, a technologist, and a parent, I firmly believe that we should strive to create AI software systems that are designed for people.
4 things we can do
It is of vital interest to all of us that AI systems are designed to maximize human potential and protect humankind’s best interests. But how? Here are four things we can do to ensure appropriate applications of AI for people.
First, we must design solutions, not in a vacuum of like-minded technologists, but instead closely with users in a collaborative, multidisciplinary, and demographically diverse environment. This is the design thinking creed that has proven itself highly effective over the past decade. Ever since Tim Brown of IDEO popularized this user-centered method in his seminal article about design thinking in 2008 in the Harvard Business Review, thousands of companies have successfully set up multi-disciplinary teams to iteratively explore and solve problems with the end user’s point of view as the engine driving every stage of the process.
Second, we should stay focused on our responsibilities to create solutions that benefit humans. These ethical considerations must not only be pondered but actively integrated into the design and development process. Keep in mind this quote from one of the greatest humans of all time, Leonardo da Vinci: “I have been impressed with the urgency of doing. Knowing is not enough; we must apply. Being willing is not enough; we must do.” Teams should formulate a checklist to help identify the biggest ethical challenges they face in their AI projects and make concrete plans on how to address them.
Third, we should strive to design interactions between humans and machines in a way that caters to people’s evolutionary heritage. Many thousands of years have given us skills and abilities to interpret and manage our physical, social, digital, and emotional environments. By adapting interactions to people’s innate capabilities, we can make working with machines delightfully natural and productive. That not only makes the interaction more fun for us, it also improves our chances of understanding what AI’s algorithms are doing.
And finally, we should consign the right tasks to machines and keep people in charge of tasks that are better off in their hands. We must carefully consider which tasks to automate and which to keep under human supervision.
Figure 1: The Process of Designing Artificial Intelligence for People
Machines are clearly superior to humans at certain tasks, but before we delegate something to them, we should ask ourselves some questions: Will the machine really be better than a human at this task? Will we benefit from this particular application of automation? What are the ethical implications?
For example, could we imagine our elderly parents in a nursing home without a single human employee? A robot could theoretically be a great caregiver to a patient with dementia. With detailed knowledge of her personal and medical history, an AI could reminisce with her about her past and prompt her with mental exercises to slow the onset of the disease. It could also tell her jokes to keep her entertained and wouldn’t mind feeding her dinner over the course of two hours if that’s what it takes to get her to ingest her daily calorie intake. One shudders to think about the level of attention given by over-worked human caregivers in many of today’s understaffed institutions. But a nursing home with just machines and robots, without a person to hold a patient’s hand and listen when he is sad? Even if we could imagine this dehumanized scenario for the elderly, what about in a kindergarten? Before automating a task, we should consider very carefully the impact on everyone involved and what role we still really want humans to play in our society.
There is really no question that AI will automate many tasks currently performed and controlled by humans. Perhaps when people can delegate tasks to machines that they couldn’t delegate before, their workday will be liberated from mindless chores, cumbersome software, and disappointing dead ends.
Imagine what opportunities will be unlocked when everyone who wants to can perform work that is highly effective, highly strategic, and highly creative. There is certainly no shortage of problems in this world, and our appetite for a better standard of living seems infinite. Imagine millions of people, supported by an ultra-intelligent system, making remarkable achievements every day. Imagine everyone empowered to contribute to solving the world’s most pressing issues or even just those things they are most passionate about. The idea is so appealing, it is almost magical.
It’s a future that I am very excited about. Yet, to make sure that the future is more utopian than dystopian, we cannot be complacent how AI develops. We need to actively discuss and design it together.
For more on the ethics of AI, read more about SAP’s guiding principles for artificial intelligence (AI).