Guest Article
Biography
Carl Wocke is the Managing Director of Merlynn Intelligence Technologies – a company on the forefront of human to machine knowledge transmission using machine learning and artificial intelligence techniques. He is considered as one of the pioneers in building technologies that enable human based reasoning within machine learning environments. Carl consults with organizations globally in areas ranging from risk management, banking, insurance, cyber crime and intelligent robotic process automation.
Artificial Intelligence And The Role Of The Human
Overview
Artificial Intelligence (AI) is upon us and indeed so much so that organizations now have little choice but to have some activity or strong opinion in this space. Ignore at your peril.
A simplistic description of AI could well be that of a technology comprising complex algorithms that are used to mine large sets of data. One of the objectives being to uncover previously hidden insights that could assist the organization in finding ways to better deliver its value proposition, optimize resources and create some competitive advantage. This then is the first version of what AI should be delivering on.
One AI becomes mainstream the natural march towards differentiation will mean that successful organizations will look to innovative approaches to harness this AI technology. Probably one of the most significant ways organizations look for differentiation is by finding larger, better quality data sources. AI is primarily fed by data, and not just any data…Big Data…the bigger the better. In theory that is.
In understanding the likely development path of AI it might be useful to look at why AI and indeed systems were created in the first place…
Consider how a business typically grows. A business starts with a product or service which is offered by a small team. The demand is small enough to allow the team to employ a high degree of manual or human content in the delivery of the product or service. The thing that makes the business different is encapsulated in the way the team deals with the customer. The secret sauce. Then along comes success and there is a need to grow and scale the business. An option is to hire more people but in most cases there is a inevitability that sees the investment into systems to assist in business growth. The systems are really designed to replicate the particular way that the business delivers its products and services and the way it treats its customers. Systems are in reality rule based and what happens as a consequence is that simplistic rules will make consequential decisions in the business. An example of which being when a debtor fails to pay their account on time. The system unemotionally freezes their account. Had the business still been small enough, then the owner may have made a different call. At best large organizations will pass exceptions like closing accounts to a team of humans who will be called on to make the right judgement call on something like freezing an account. Passing exceptions to humans is unfortunately and in reality just creating a bottleneck in a process. Take a bank for example where the bank monitors for anti money laundering activities. Transactions are flagged and then passed to expert assessors to validate or invalidate the transaction. A massive bottleneck.
Like the banking example there are complex priorities and complicated operational decisions in Retail, Utilities, Mining and other industries where bottlenecks are created through exceptions being passed on to experts. Where non-linear, complex decisions require subjective judgment experts are called in to make a call and developing expertise takes knowledge, skill, and time. Imagine a Call Centre example where organizations always have top performers who exhibit an innate ability to make decisions faster and with better outcomes than their peers without exceptions being referred to other experts in turn.
The next major evolution in AI is therefore very likely to be the ability for AI to acknowledge additional fuel sources. That of the human expert. The ability to not only learn from data but to learn from people…from experts. Systems that think like people.
In the banking example, there is bottleneck around clearing a flagged transaction. AI can be used to mimic the top assessor thereby making the same judgement call that the assessor would have made. AI used to scale expertise vs replacing expertise.
AI should, and most likely will be used to scale an essential resource called expertise and specifically expertise within decisioning. The nature of expertise is something that is not readily definable or even found within data. If it was then it would not be expertise. It would be common knowledge. Expertise is an instinct, a gut feel. Just knowing the right call to make.
AI is the right technology approach to capture this valuable resource. A likely consequence to this process will be the ability to capture instinct emotions like empathy and ethics.
There are a number of doomsday perspectives on the future of AI and the threat to our very existence. There surely is risk when considering computational capacity replacing highly repetitive tasks. However, when it comes to abstract thinking, intuition and empathy and the added reality that AI will mimic vs replace, then surely we could well be looking at the next generation of AI as being humanistic rather than robotic.
Interesting link for further reading:
http://www.itweb.co.za/index.php?option=com_content&view=article&id=162082