As intelligent systems become increasingly infused within business, it’s fair to ask: are we still in charge?
The first wave of enterprise digital transformations that began roughly 10 years ago focused on building social, mobile, analytics and cloud capabilities. Organizations learned how to design, deliver and activate connected business operations, connected services, connected products, connected teams, connected content and connected experiences.
Even now, most companies have not completed their digital transformation and remain on the journey to credibly become digital. Meanwhile, the next wave of digital has arrived, largely driven by the promise of artificial intelligence (AI).
The art of the possible
Companies are racing to capitalize on a new art of the possible, inspired by the next generation of automated systems (commonly referred to as machines, robots, bots, chatbots, autonomous systems, etc.).
In the mainstream media, the most frequent headlines are about AI these days, but the automation narrative is far broader. It encompasses a spectrum of computational technologies that perform tasks previously conducted by humans. These range from software “bots” that automate manual processes and autonomous delivery drones to cognitive procurement systems and virtual HR advisors.
At the intersection of three primary, adjacent domains (advanced analytics, robotics and AI), automation is playing out at scale on many fronts. We are automating trust (blockchain); insight (machine learning); awareness (IoT); conversation (natural language processing); authorship (natural language generation); recognition (computer vision); and so on.
There is also widespread confusion, and intrigue, around the idea of an automated system that can see, talk, compose and otherwise appear “human” in many ways. But when you strip away all the fancy rhetoric, AI is just simple math executed on an enormous scale.
The more calculations a system can process, the more possible it is for that system to emulate human-like cognitive abilities. With the advent of cloud infrastructure, GPU-accelerated processing and deep learning architectures, it is now commercially viable to perform this math at such speeds and efficiency that AI, with its human-like cognitive abilities, can be embedded directly into business operations, platform architectures, business services and customer experiences.
Who’s in charge?
There have been daunting headlines on the issue of “who’s in charge.” A global e-commerce player recently introduced a physical retail experience that runs almost entirely without people, enabling customers to shop and check out without stopping or swiping any credit cards, thanks to a network of sensors and AI.
An August 2017 article in Business Insider titled, “JP Morgan takes AI to the next level,” detailed how it might be the first major financial institution to apply AI to real-time trade execution as well as client management. The legal field is booming with AI startups, and a major law firm hired the first AI lawyer, ROSS.
As the algorithms on which AI depends become infused into almost every aspect of business, it’s perhaps no surprise that regulators are beginning to take a close interest in them, with a view to protecting both customers and employees.
Given the growing importance of algorithms and automation, it might be time to appoint a chief automation officer to the executive team and to update risk management models to monitor and mitigate the potential impact for algorithmic bias on business operations and customer experiences.
A tool for empowerment
The fear that machines will not only replace people at work, but “take over,” is rooted in a movie-type view of AI and can be rebutted.
It is true that many activities long considered the exclusive domain of human intelligence have been surpassed by AI, and much faster than experts anticipated. This includes, the ability to read lips, identify lying, interpret human emotions based on facial expressions, and make medical diagnoses.
Human intelligence, however, is not singular; it is combinatorial, reflecting a suite of cognitive and sensory abilities working together. As “smart” as we may perceive an AI, based on the sheer speed and complexity of its information-processing abilities, it can only approximate human intelligence in very narrow applications.
More importantly, AI cannot independently and autonomously modify itself, or any components of its architecture, nor can it expand the scope of its intelligence and apply its current domain knowledge to another domain.
For example, an AI trained to play chess can be retrained to learn checkers, but based on its knowledge of chess alone, it would fail to know how to make the very first move at checkers. AI is a tool for empowerment, and every component of its design and functionality depends on the ingenuity, engineering and supervision of humans.
Second, machines automate tasks, not jobs. Some jobs are more “automatable” than others, no question, and the nature of employment will be reshaped by these technologies. Nonetheless, the increased use of automated systems at work, and the integration of people and machines whether on factory floors, in restaurants or at call centers, will lead to new job categories and many other benefits, such as increased employee productivity, more meaningful work experiences and safer work practices.
To understand the latter point, consider the pipeline inspector for an energy company, who previously relied on helicopters but who can now replace these with unmanned aerial drones to survey facilities and properties faster, cheaper and without putting humans in harm’s way.
New paradigms of performance
The full scope of impact that we can expect from AI is defined entirely by the designers of AI. If we decide to transfer more control over business processes and customer experiences to an automated system, then we should be prepared to accept the consequences of that shift. However, the burden of designing, training, curating and monitoring those algorithms remains squarely on people.
Ultimately, as humans and machines collaborate at work, it should drive new paradigms of operational performance and shareholder value, with people still very much in charge.
