Over the next few years artificial intelligence technologies, also known as cognitive technologies, will likely have a profound impact on work, workers, and organizations. These technologies can and will be used to eliminate jobs. But they will also make it possible to redesign work, creating new opportunities for workers and greater value for businesses and their customers. To prepare for the future, business leaders should understand their four main automation choices and the cost and value strategies described below.
After analyzing over 100 applications of cognitive technologies, my colleagues and I found that these applications tend to fall into three categories: product, process, and insight. Product applications embed cognitive technologies in products to provide "intelligent" behavior, natural interfaces (such speech and visual) and automation. Process applications use cognitive technologies to enhance, scale, or automate business processes. Insight applications use cognitive technologies to reveal patterns in data, make predictions, and guide more effective actions.
Each category of application has distinct impacts on work and workers. To create products with embedded cognitive technologies (such as robotic vacuum cleaners or televisions you can control by voice) companies need workers with cutting-edge technical skills. They also need user experience designers conversant in the possibilities presented by cognitive technologies such as interfaces, learning, and intelligent automation.
Process applications may ease, change, or eliminate some kinds of jobs, as cognitive technologies make it possible to automate certain tasks such as those involving the processing of unstructured text, speech, and handwriting.
Insight applications tend to change the way a job is done--often enabling it to be done more effectively. In one example we know of, a major technology company has begun to machine learning to reveal insights about the buying behaviors of its customers. Their salespeople, rather than relying solely on their own judgement, now take direction from a software application that recommends which customers to call next and what to offer them. They still have their jobs, but are doing them more productively--their sales have risen as a result of following the advice of the machine learning application.
Since cognitive technologies make it possible to automate many kinds of tasks that could not be automated before, it is worth considering what we have learned about automation over the years since it first started appearing in factories. Although automation is undeniably valuable, decades of research have shown that it does not always deliver the foreseen benefits and can have unintended consequences.
For instance, automation is sometimes seen as a way to avoid the errors that we imperfect humans inevitably make. But automated systems are created by humans and inevitably have their own flaws. Further, research has shown that people have a very difficult time monitoring automated processes--they can't stay focused when there is little for them to notice or do. "Cognitive underload"--insufficient stimulation and engagement from excessive automation--has been found to reduce worker performance on a range of tasks. Automation has been found to cause workers' skills to atrophy. Badly designed automation may even undermine our identities and sense of self-worth, as technology commentator Nicholas Carr has argued.
Four approaches to automation
- Replace. In this approach, technology is used to perform a job that used to be the primary activity of a person. Examples: replacing bank tellers with automated teller machines or call center workers with interactive voice response systems.
- Atomize/automate. This approach involves atomizing, or breaking up, a job into pieces, and automating as much as possible, leaving humans to do the non-automatable pieces and possibly supervise the automation. The transition of artisans to assembly line workers is an example of this. Relying on machine language translation and leaving professionals to "clean up" the results is another.
- Relieve. In this approach, technology takes over tasks that workers don't relish or are overqualified for so they can apply their skills to more valuable, more interesting work. The use by Associated Press of machines to write routine corporate earnings stories so journalists can focus on in-depth reporting is an example of this.
- Empower. In this approach, technology makes workers more effective by assisting them and complementing their skills. An example of this approach is IBM Watson for Oncology, which recommends cancer treatments to physicians, citing evidence and a confidence score for each recommendation, with the goal of enabling physicians to make more fully informed decisions than they might have been able to before.
Neither the type of job nor the technology used to automate it determines which automation approach to follow. This is a choice to be made by systems designers and, even more importantly, leaders and strategists.
Cognitive technologies present another choice to business leaders and strategists: use them to cut costs, or use them to create more value. Cost-cutting strategies may point toward replacing people with machines, whereas value strategies may point toward empowering them to do higher quality work.
The choices we make about how to apply cognitive technologies will determine whether workers are marginalized or empowered, and whether our organizations are creating value or merely cutting costs. There is no single right set of choices for organizations to make. As leaders prepare to bring cognitive technologies they should be aware that the choice is theirs.
Read the full article from which this essay was adapted here.