Artificial Intelligence – Fake News or Alternative Facts?

One of the most worrying developments in 2016 was the rise of fake news. And from the outset of 2017 it appears distortion of reality will become the norm. Perhaps we shouldn’t be so surprised; the marketing industry has always been about changing perceptions, and the technology industry in particular is notorious for generating hype. There are countless examples of hyperbole about new technologies that in the end fail to realize the optimistic predictions. The latest new kid on the block is AI, the development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making and language translation which of course also encompasses the whole field of robotics. The rise of robotics and AI has provided copy writers with great material. A Japanese life insurance firm recently announced it is laying off 34 office workers by the end of March as a result of the implementation of an AI based claims adjudication system based on IBM’s Watson Explorer. The system is described as reading tens of thousands of medical certificates, factors in length of hospital stay, medical history and procedures before calculating payouts. This experience clearly supports the 2016 Nomura Research Institute report which forecast that nearly half of all jobs in Japan could be performed by robots by 2035.

Sean Welsh, researcher in robot ethics at the University of Canterbury, New Zealand goes even further making the prediction that robots will replace many humans: perhaps most humans in the workplace. Oxford Business School writers Carl Frey and Michael Osborne predict massive technological unemployment over the next 20 to 30 years. Something like half to three-quarters of existing human jobs are “vulnerable to automation” they say. Martin Ford, author of a popular best-selling book, The Rise of the Robots, predicts an explosion in robotics; his rationale is based on the emergence of the open source robot operating system (ROS) conceived by the Stanford University’s Artificial Intelligence Laboratory, will make the technology ubiquitous.
Last year Professor Stephen Hawkin, one of Britain’s outstanding scientists told the BBC that, “The development of full artificial intelligence could spell the end of the human race.” This view has been supported by Microsoft founder Bill Gates and SpaceX and Tesla founder Elon Musk.
And these examples are a very small selection of the vast amount of media coverage of this technology. With a basic message that is unmistakably threatening peoples jobs and livelihood, many people are not surprisingly very concerned and this reinforces the widespread public opinion in the USA and United Kingdom that in 2016 pushed back against the inevitability of globalization.

So what’s the reality. Will artificial intelligence really eliminate millions of human jobs? Let’s look in just a little more depth at the case study of the Japanese life insurance company and its use of IBM’s AI technology. Over the past few years many insurance companies have been modernizing their claims adjudication systems. For many the trigger for change has been that the existing claims systems are based on mainframe technology, often decades old, and that once elegant software systems have become impossibly complicated by constant change of regulations, change of health and insurance industry standards, new product and plan introductions and the continual mergers and acquisitions that involve systems rationalization. The result is the firms’ ability to respond in a timely manner at reasonable cost to continuing change is dramatically reduced. And just to make the situation even worse, the technology used to build the systems is decades out of date, and the original support staff are now retiring, and younger personnel are not interested in using the older technologies.

The way AI system works is that it searches for similar patterns that have occurred in the past and compares current cases, and where they are common can recommend decisions with minimal or no human intervention. But the AI system requires a vocabulary that describes the claims business policies – and in most firms this is locked into the existing software system, all muddled up with design and computer language specifics. In larger insurance firms it typically requires teams of business analysts working for multiple years to reconstruct the complexity of the firm’s claims policy in a well-structured manner that can be used to guide claims adjudication.

For most insurance firms therefore this requires a significant modernization project. The systems often need to be redeveloped to use the rationalized policy base and AI based decisions. From this description it will be seen that there’s a massive amount of work going on to make all this happen. Existing systems development personnel are typically not skilled in the new development or AI technologies, but they are expert in the existing claims business policy and are commonly employed in developing and managing the business rules base. Similarly claims adjudicators are commonly subject matter experts and play a critical role in these tasks as well as quality assurance, systems testing and governance. The reality is that when the business rules knowledgebase and AI system are operational and mature the number of human adjudicators actually making decisions on claims will reduce. But there are many new roles in the redevelopment and subsequent operational phases that mitigate the overall reduction in headcount.

So returning to the Japanese case study we might speculate why the press release focused on job losses. First industry experience is that modernization efforts like this are undertaken to reduce and govern claim pay-outs. Labour costs of the operational system are less important. Also, while the press release concentrated on the AI element of the adjudication system, it is quite likely that the company would have been involved in a major effort such as described here, at least to rationalize the business rules involving many people over multiple years to re-establish the existing claims system, and retraining of staff to establish the business knowledge bases and preserve the deep skills and expertise. The case study reported increased productivity of 30%, ROI in two years, cost savings of £1m per year all set against a system cost of £1.4m. These costs seem trivial compared to the real cost of development that most firms in this situation would need to spend.

I note IBM CEO Ginni Rometty spoke last week at Davos, and delivered a clear message to business and political leaders that artificial intelligence will not lead to mass redundancies any time soon. Similarly, in an interview last week Microsoft Chief Executive Officer Satya Nadella said, “Microsoft Corp. and its competitors should eschew artificial intelligence systems that replace people instead of maximizing their time.”

New technologies always cause change in skills and roles. This is entirely natural. However, experience shows that as roles evolve and disappear, new roles emerge. In the early stages of any new technology it’s often difficult to envisage what those new roles will be. However, as discussed experience to date shows that the new technologies always require massive efforts to realize. And skills development and lifelong learning is going to be even more important than it is today. Clearly leaders in some tech companies have got the message that it’s not a good idea to promote their technology as job elimination. Let’s hope the media will reflect that message.

About davidsprott

Artist, writer, veteran IT professional
This entry was posted in Artificial Intelligence and tagged . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s