Machines Want to Help You Do an Amazing Job
Have you been alarmed by reports of artificial intelligence (AI) stealing your job? Given the attention the media has given to this supposed threat to your ability to pay your bills, it’s only human to feel some concern about the rise of the machines.
But let’s not get carried away. This is a rise that’s been taking place since the invention of the wheel. The very point of a machine is something that reduces the amount of effort that we must produce ourselves to solve a problem. As elementary students in science class, we were taught about simple machines — pulleys, levers and ramps — that reduce the work required to raise a load, organize objects, direct actions, and myriad other tasks. While modern machines are much more sophisticated, they still require human supervision. The job the human performs has evolved, but it has not disappeared.
While intelligent machines can and will replace certain job functions, the concept of AI was never based on a fundamental human vs. machine battle. As with any machine, AI was conceived to help humans save effort when solving problems. As machines take on more tasks, as they have done for centuries, so too will our experience of evolution in the jobs we do as we take advantage of ever-increasing productivity.
Cognitive computing is your machine assistant
There is a difference between AI machines and the mechanical ones that were the targets for job loss anxieties of the past. Cognitive computing, a subset of AI, mimics the thought processes of the human brain, including association, prediction, correlation, and inference. These skills, once thought to be uniquely human, mean that jobs that require brain rather than brawn can now be performed by a machine. This is significant, hence the headlines, but as with previous evolutions we’ll see jobs change rather than being lost.
Like a human brain, a cognitive computing system is able learn over time. It can compare past outcomes and refine its course of action. Unlike a human brain, cognitive computing can draw on the computational power of an unlimited number of processors to analyze vast quantities of data. As a result, this technology has the ability to improve the efficiency and accuracy of knowledge workers — experts whose job functions are to review, analyze or use information.
Cognitive computing is already working to support humans in the financial, government, and healthcare sectors — all of which are awash in data. Organizations in these fields have been pioneers in implementing cognitive computing to power intelligent assistants — task-specific support that amplifies human skills to produce the best possible outcome.
This is how many of us will experience AI at work. While some AI systems will operate autonomously to execute specific tasks, knowledge workers will use cognitive computing systems to make better sense of vast and complex information resources. The machines will be our assistant in finding the insights that enable us to determine the best course of action.
Machines are helping to end human trafficking
There are already many remarkable examples of cognitive computing in action alongside humans. One of the most heartening is found in law enforcement. This is a field where the need to make thorough and accurate assessments is put under huge pressure by the time-critical need to act. With its ability to find vital insights within huge volumes of data, cognitive computing is highly effective at general surveillance and investigation tasks. This has particular value when used to unravel black market businesses such as human trafficking.
In our digital age, it can be remarkably difficult to identify traffickers and their victims because the evidence bound up in digital records can disappear in a matter of days. According to Thorn, an anti-trafficking organization, over 75% of child prostitutes are advertised online. This equates to 150,000 ads per week — an excellent source of evidence if it were not that the ads typically disappear within six days or less. With evidence evaporating faster than it can be analyzed by humans, the ability for law enforcement to leverage technology is vital to the success of anti-trafficking programs. Cognitive computing is used to identify victims and gain insights about trafficking rings in a matter of minutes, rather than days, helping officers act faster to assist victims and be more successful with prosecutions.
Giving time to patients, not paperwork
Labor represents more than 50% of healthcare costs. A large proportion of that time is spent creating, processing and analyzing clinical documentation. In fact, for every hour of physician patient care, two hours are spent reporting. Experts predict that the introduction of increasingly complex regulatory requirements will increase the reporting burden, meaning even less time for face-to-face clinical interaction. Machines that can take on some of the reporting burden will help doctors and nurses redress the balance and focus on patient care.
In some hospitals, clinicians are experimenting with cognitive computing to improve care quality. The technology can be trained to identify evidence of disease within a patient’s electronic health record or radiology report. For example, a pathology report can be analyzed within minutes of its completion, alerting a physician or nurse about whether a patient has indications of cancer. Cognitive computing assists healthcare knowledge workers by focusing on the patients who need follow-up care, bringing help to those who need it more quickly. This approach not only raises quality but also improves increase productivity. Care professionals can spend more time with patients, improving care outcomes, but without adding to costs.
Helping banks stick to the rules
Compliance departments at financial institutions face a daunting task to inspect the millions of electronic communication interactions that take place within their organization each day. Prosecutions for market fixing and insider trading have produced evidence of wrongdoing in emails and instant messaging systems. To defend their institution and identify culprits in the act, compliance officers are extending their gaze to encompass phone calls, images, documents, and any other form of electronic communication.
In large banks there are often thousands employed at this tasks, but even so it is simply not feasible for human beings to analyze the millions of messages produced on a daily basis, which, in turn, makes it impossible to conduct a quality review on a consistent basis. Cognitive computing has been transformative in helping financial knowledge workers find the rare instances of illicit activity in the sea of normal communications.
Organizations like UBS and Point72 use cognitive computing technology that not only flags suspicious language, but also makes sense of the context of conversations to reveal hidden meanings. This ability to augment human skills makes it possible to identify communication patterns that commonly indicate to fraudulent activity such as insider trading, even when people are careful disguise their intentions. This acuity enables compliance officers to be more effective at finding incidences of noncompliance and helps them learn the tricks used to commit financial fraud. Mercifully, it also reduces the number of false alerts.
The fundamentals of our future
These examples of domain-specific AI describe an emerging future where we work with machines as partners. The difference is that they will play an increasingly active role, which will help us to achieve better outcomes. Comparisons of cognitive analytics and a human analyst have striking similarities, but the reality is that aligning machine and human intelligence to expand our goals is evolutionary, not revolutionary.
Of course, it doesn’t end here. Computers will get smarter, algorithms will get longer and we’ll see real advancements within AI itself as domain-specific augmentation turns into something more adaptable. Eventually we’ll develop ‘artificial general intelligence’ — machines able to pass the Turing Test that can take on multiple and broad tasks.
This will be a disruptive process, but humans should not fret about being replaced by our own inventions. As we embrace new forms of technology, we grant ourselves new opportunities to spend more time on problems and challenges that will undoubtedly remain. Our progress is guided by our humanity and the intelligent machines that we use to achieve it are fundamentally no different than those of any period from our past.