Decision making remains a complex discipline
Whilst it’s true that computers are able to do things at greater speed and with more reliability, artificial intelligence provides the next layer of capability to allow computers to continue this role. That said, we should still remember that computers do not program themselves (well, not yet anyway….). Decision making is a process that requires both elements of data gathering and data analysis. Computers have been good data gathering, and within bounds that are set by programmers, are becoming increasingly capable providing data analysis and cognitive insight.
Machines are more consistent in their analysis
By now, we’re familiar with the epithet that ‘Computer says ‘No’’ has taken over, where the users of a system simply defer the decision making to the system and have no idea sometimes why it came up with the decision that it did, they simply go with it. What’s overlooked in this simplistic interpretation is that the system was provided with a set of rules to operate in the first place, and the fact that some of the users have perhaps not been trained adequately to understand the results, or can’t be trusted to apply interpretation to the results themselves is not ‘the fault’ of the system. The system, after all, has simply applied the rules.
However, this is where the balance of this question lies. Within designated bounds, we are giving the computer more scope to establish what looks like its own decision-making process, largely because of the speed at which it works, and also because we can rely on the computer to apply the rules it has been provided, but it doesn’t come up with a new rules base on its own. Even if a machine learning app throws up an insight that was unexpected, it has only done so within the confines of the learned experience that forms its knowledge base.
Machines don’t get tired or bored
Machine learning and cognitive insight applications enable solutions use predictive rules that for example automatically recognise anomalies in data sets to support roles performed by operators, not replace them as such. I have in mind labour intensive data recording for quality control of production processes. It’s not that the humans don’t know how to do it, or what the implications may be, but having the machine record data and tell you when readings are going out of tolerance is a more effective use of the time of a Quality Engineer.
The machine is helpful in that it doesn’t get tired, and potentially miss significant data or events that arise. It might be that in doing so, it is undertaking 99% of the data recording that used to be done by operators, but it is also freeing up those operators who might otherwise become tired or bored with repetitive processes that they miss significant data or events. This provides the operator with an ability to concentrate on a greater number of machines, knowing that the computer is undertaking the monitoring function, which makes for better, more effective processes whilst allowing the operator to undertake functions that the machine cannot do. People can do a better job as a result.
Machines Don’t Have Intuition
Machine Learning is valuable because of its ability to generate unlooked-for cognitive insight, or one that might otherwise have been overlooked or discarded because the human deemed it irrelevant and discounted it. Where businesses have in the past relied on the experience of their staff to recognise a situation, this can be unreliable and inconsistent. A particular set of circumstances may never have been experienced previously, but if those circumstances can be programmed into the system, it can respond to learn from experience and assist appropriately.
Biases and blind spots can still exist in Big Data, but the interpretation of them is likely to be more consistent. In that sense, it’s because the machines don’t have ‘intuition’ in the way that people do that their capabilities can be applied more consistently to achieve parts of the function where their success is useful. It’s consistency of decision making that can increase consistency in the process, reliability in outcome and improvement in business performance.
Making the most of Big Data requires both greater analytics capability and greater ability to put cognitive insight to productive use in meaningful business decisions. Neither an ‘all analytics’ nor an ‘all intuition’ approach will deliver the goods. The greatest performance gains are achieved not when machines are used to replace employees, but when they are deployed to work alongside them. In such collaborative relationships, people help machines become better, and machines enable people to achieve step-level increases in performance.
If you wish to share your thoughts on this article, you are welcome to comment below, or to contact the author through the Contact Form or by phone on +44 113 242 3795.