AI is integrating a long list of some of the most innovative and disruptive technologies available: machine learning; cognitive and computer vision; conversational capabilities; human-to-machine user interfaces; predictive data analytics; cybersecurity; IoT and intelligent monitoring.
It is a huge range of applications that will take considerable skill to implement if they are to be optimised in work environments around the world. Despite all the advances in management software and techniques, the most effective path to extracting the maximum value from these technologies is through convergence with robotic process automation (RPA).
Robotic process automation
This proven technology is already bringing together AI and cognitive tools to achieve productivity gains of several orders of magnitude whilst simultaneously slashing costs. While RPA streamlines workflows and performs routine administrative tasks at high volume in a fraction of the time it takes humans and to far higher standards of accuracy, the addition of AI and cognitive capabilities promises smarter business decisions and the potential for much more far-reaching positive change.
The ability to learn is the distinctive characteristic of AI and it means that computers are no longer limited to simply carrying out instructions written by humans. Instead, they can continuously learn from new data and perform better than their human programmers. It is what is described as moving from rules-based to experience-based operations.
AI will take over countless tasks such as driving cars, diagnosing diseases or providing customer support. AI’s superhuman performance will generate massive increases in productivity. A June 2017 study by consulting firm PwC, predicts AI will stimulate $15.7 trillion in additional wealth for the world by 2030.
While AI can be considered the “thinking” component of the digital workforce, RPA embodies the component that “acts”. Computer vision and the crunching of all types of data (structured or unstructured) embodies the “sensing” component. All of this data will be then fed into the AI solution, completing the circle to provide the power to RPA software automation which can work attended under the direction of a human worker or unattended and completely autonomous.
Building trust in AI may take time
Trust, however, becomes a factor here because within every organisation there is likely to develop a conflict between retaining control of the AI and trusting its autonomy. An employee can be trusted to perform as expected through training and personal performance controls. But how do humans trust a robot or any automated process, for that matter, to perform as expected?
It is important to recognise the difference between an RPA robot, and a robot as an AI solution. While the RPA robot is programmed to perform certain tasks, the AI solution is trained by observation.
Trusting AI machines will take time and requires transparency and analytics from the point of implementation. Finding a transparent way of helping humans see how the machines learn and use this knowledge in the correct way and take the right decisions will not be easy.
Fear of change and lack of personal employee engagement can also have negative impact on the successful adoption of AI-powered automation. It is important for all organisations implementing RPA with AI capabilities to educate their workforces. Artificial Intelligence should be viewed as a technology that makes each employee more productive and opens up access to new capabilities. Training is necessary but should not take too long or become intimidatingly complex.
The wider effects of automation
There are limits to the abilities of today’s AI which have other ramifications for the relationship between society and automation.
Repetitive tasks, such as auditing or insurance and credit administration are likely to disappear and many of the basic information-accessing and processing aspects of the professions such as medicine and the law are likely to be performed by AI and cognitive applications. Yet any role that definitely requires complex human interaction or the management of people is likely to be retained and indeed increased. These jobs are beyond AI’s capabilities in terms of creativity, strategy-formulation, sociability and care. Although the technology goes way beyond human capabilities in the digital sphere it lacks social skills or empathy.
The concern here is that AI will radically alter many middle-ranking jobs that involve a degree of administration and information analysis. Roles that will remain relatively insulated will be at opposite ends of the income spectrum – from CEOs to carers, podiatrists and hairstylists.
As a result, there has been much discussion about providing a universal basic income to citizens who have lost their jobs to technology. This is to miss the opportunity presented by this transformative technology.
Such a revolution in how we relate to work will require a rethink from all corners of society. In the private sector, instead of simply viewing AI as a means for cost-cutting through automation, businesses will create new jobs by seeking out symbiosis between AI and the human touch. This will be especially effective in areas such as healthcare and education, where AI can produce crucial insights but only humans can deliver them with care and compassion.
We are already seeing that in many organisations where AI-powered ‘Smart’ RPA is implemented, there are very low levels of job losses as the humans are freed from the mundane to move up the value chain; i.e., employees have more time to devote to profitable, creative tasks which increases their engagement.
A rethink may be required
There is no denying that some roles will disappear and that will require fresh thinking about how to use the wealth generated by AI to reorient our economies.
Some of this wealth could be used to support socially useful work such as care or charity work, parenting or home schooling of young children, assisting aging parents or helping a friend with mental or physical disabilities live life to the full.
By requiring some social contribution, we would foster a public ethos that focuses on our usefulness to one another and wider humanity. Many difficult questions remain to be answered, of course, before we could consider implementing such a sweeping and idealistic policy.
But in the meantime, it is important that business concentrates on obtaining the huge gains in productivity and efficiency from the most effective implementation of RPA and AI. This is the one certain way of harnessing the combined potential of thinking machines and humans.
Guy Kirkwood, Chief Evangelist at UiPath