While examining the potential of AI, McKinsey & Company identifies five important factors to take into consideration: “(1) technical feasibility; (2) costs to automate; (3) relative scarcity, skills, and cost of workers who might otherwise do the activity; (4) benefits (e.g., superior performance) of Automation beyond labor-cost substitution; and (5) regulatory and social-acceptance considerations.” Although all of these factors are considered when deploying AI, in this article, we solely focus on the issue of social acceptance.
Given all the hype surrounding AI, ML, and Automation in general, it’s important to take a step back and consider which tasks, if any, can only be done by humans. No matter how many data points we enter into a Machine Learning platform, there is a certain type of intelligence that is uniquely human.
In a New York Times op-ed piece, MIT professor and Psychologist Sherry Turkle wrote, “Technologists presented us with Artificial Intelligence, and in the end, it made us look differently, and more critically, at the kind of intelligence that only people have.”
Empathy Cannot Be Replicated
Computers will never be able to experience pain, nor will they be able to fear death or experience real empathy. Although computers will certainly be able to feign empathy, they will never be able to actually possess it.
Cognitive empathy (the ability to recognize other humans’ mental states), as well as emotional empathy (the desire to respond to other humans’ mental states with appropriate emotions), are innately human conditions.
Although most of a person’s empathy is developed during their childhood, a recent study from scholars at the University of Cambridge found that at least 10 percent of individual differences in empathy can be attributed to genetics. This doesn’t apply to mentally handicapped individuals or those with personality disorders, but for the sake of this argument, we’ll focus on the average human.
Technology’s Assault on Empathy
In her 2015 book Reclaiming Conversation, Turkle hyperbolically argues that “technology is implicated in an assault on empathy.” Given the extremely fast clip at which technology is moving, it makes sense to give this accusation some attention.
Despite the fact that humans have an innately different form of intelligence, we are actively creating machines that offer artificial empathic responses. As an example, Google’s human-sounding AI, Duplex, utilizes hesitation markers, false starts, and fillers, including “um,” “ah,” and “you know,” to better mimic the cadence of human speech. So, the question really becomes, “Do we find it socially acceptable to engage with human-sounding AIs? “or “Do we even care if we know whether it’s a real person we’re speaking to?” The answers to these questions likely depend on the generation in which one was born.
A New Generation Is Emerging
Turkle argues that the creation of machines that offer artificial empathy dilutes the human experience. She posits that “children will lose the ability to have empathy if they relate too consistently with objects that cannot form empathic ties.” Due to a constant influx of emerging technologies, young people may indeed be losing the capacity for empathy, as they increasingly rely on Human-Computer interaction. If this is the case, then today’s toddlers —who are growing up with AI embedded in their daily lives—will have less of a need for empathy and human contact.
However, from a business perspective, this is encouraging on many fronts. As more people become comfortable with Artificial Intelligence embedded in applications, the more popular it will become until we reach a tipping point, in which we can’t decipher whether the entity on the other end of the line is a human or not. But will people find such widespread adoption socially acceptable? Again, the answer will likely be dependent on the generation in which one was born. If Turkle is correct, today’s toddlers will embrace AI adoption later in life, as they won’t require as much authentic empathic interaction.
Where AI Will and Won’t Work
There are clear benefits to the adoption of AI for intensive, routine and error-prone tasks where empathy is not a factor. For instance, AI has already transformed Data Collection and Data Processing, as the appropriate data points are gathered, sorted, and analyzed before passing on to a human. We’ve seen AI handle an astronomical amount of data in a short period of time, saving humans countless hours of work. AI has also been effective at automating routine tasks, such as password resets for IT departments, and other core business operations including Sales Forecasting, Customer Segmentation, Dynamic Pricing, Content Curation, and the timing and delivery of Email Marketing Campaigns.
That said, some software operations can never get automated. Within certain industries, especially health care, humans will generally want to interact with other humans. As an example, an IT manager at an organ procurement organization recently lamented to me that automation works wonderfully for password resets; however, when it comes to informing clients about their organ donation status, patients will always demand to speak with another human being rather than informing by an AI-driven software.
Moreover, across all industries, it will probably be quite a long time before sales and support roles can be fully automated, given the fact that these jobs are generally based on cultivating relationships and harnessing authentic empathy. However, if Turkle’s theory proves to be prescient, upcoming generations will be more accustomed to human-computer interaction and less reliant on empathic interaction, which will provide even more automation opportunities for software companies.