Artificial Intelligence (AI) is part of our daily lives. It is everywhere and we cannot avoid it anymore. Websites such as Amazon or eBay use deep learning techniques to profile us based on what we search and what we buy in order to suggest new products aligned with our preferences.
Streaming services like Netflix or YouTube do the same. They profile us based on what we watch, for how long we watch it and what we “like” so they can suggest new content to keep us using their platform. Google uses all the information it collects from us and applies state-of-the-art machine learning techniques to present tailored search results that are more likely to contain what we are looking for.
Biomedical domains, including healthcare, have been using AI for decades. An example of that is the use of case-based reasoning to help medical practitioners to identify illnesses based on symptoms.
Automated rostering and scheduling tools also use AI to assist nurses in hospital wards, carers and nurses in care homes, and with home healthcare coordination. Such rostering systems use another facet of AI, known as optimisation algorithms, which allow coordinators and managers to achieve much-improved efficiency when scheduling and routing staff. This efficiency comes in terms of reduced operational costs, best utilisation of staff time, improved patient satisfaction and reduced time spent on rostering and scheduling to a fraction of what is required to do it manually.
Larger corporations, such as hospitals and big logistics companies, are adopting solutions based on AI for decades. These solutions previously required huge investments in development and specialists to maintain it afterwards, but these solutions are now attainable for smaller companies.
Try before you buy
Despite this new opportunity being within reach, there is still some resistance towards embracing it. Why?
Adopting an AI solution involves much more than simply buying a product off the shelf and using it. These solutions automate methods or provide information and guidance that often demand changes in business processes, decision-making, or both.
Take, for example, the aforementioned rostering systems. A whole new business process must be built around using automated rostering to derive the fullest potential. However, before working on new processes, companies typically trial the solutions first to make sure that the change in processes works for them. But without the processes in place, the trial is compromised. So, what to do?
To answer that question, we must first understand the underlying reason behind any reluctance to adapt processes to adopt new technology. Fortunately, that is simple…
It is part of human nature to not trust what we do not understand or take for granted. We don’t need to be mechanical engineers to trust cars because we take them for granted, but back in the 1910s you would find people arguing for the use of animal-powered transportations over the use of the internal combustion engine. It took many daring people to start using the latter and prove the benefits to get the masses onboard .
But the burden cannot fall on users alone. It is our responsibility, as AI designers, to provide solutions bundled with the right tools to allow users to understand why the decisions are made. Since AI is not taken for granted at this moment, it is up to us to provide means of improving its understanding. For example, why certain rotas for staff are the most efficient, or why certain diagnostics are the most probable.
Change is here
However, from a user’s perspective, there will always be a risk. The most successful individuals in history had a moment where they had to take calculated risks, a ‘leap of faith’, and try something bold and new. Bill Gates bought what would become DOS; Elon Musk gave up his PhD to venture into the Silicon Valley. Employing a software that will automate complex tasks or interfere with decision-making can be scary at first, but such is the state of the modern AI field that the benefits are guaranteed.
The hard, experimentation stage of the field of AI is now in the past. The change is here and now. And we all know what happens when we resist change. How many businesses survived without accepting credit and debit cards? How many businesses will survive without going mobile? And now, how many businesses will survive without employing AI?
Rodrigo is Head of Optimisation at workforce management software company, Webroster. Previously the KTP Research Associate in a collaborative project between the University of Nottingham and Webroster, Rodrigo was also formerly artificial intelligence lecturer at the State University of Maringa, Brazil. Rodrigo concluded his Ph.D. studies within the Automated Scheduling and Planning Group at the University of Nottingham. He is experienced with combinatorial optimisation, including workforce scheduling and routing problems, travelling salesman problems and graph planarisation problems.
