As we heard ahead of this year's British Science Festival, widespread acceptance of automated technologies could be held back due to public fear. Even Prince Charles has been vocal about his concerns of a world dominated by machines. Despite this skepticism in the business world AI is no longer just futuristic. Although it is still emerging AI is already enabling companies to rapidly change work processes to make them more fluid or completely reimagine them altogether. To move past the speculation and amplify human work in a tangible way, businesses must put human intuition in the middle of data analytics and advanced algorithms – we call this "augmented intelligence".
The evolution of "generative" analytics
Analytics technology has moved on significantly since the passive systems of old that restricted users to limited, pre-canned data reports. Even the next iteration of systems, which empowered business users with data visualisation, did not provide the means to fully explore the plethora of data in front of them. Only a select few solutions had the qualities that allowed users to see the whole story living in their systems.
And this is where automation joins the party in a feasible way. This third wave, deemed by some as "generative" data analytics, evolves the previous notion of data discovery by automating analytics workflows with machine learning (ML). As computing power expands and the use of advanced algorithms becomes more commonplace, business intelligence systems will auto-generate analysis and statistically significant insights for business users. Therefore, the system understands business needs in the form of questions and finds corresponding analysis to help with this. As automated analytics becomes more prevalent, business decision making will increasingly be based on pattern recognition and advisory services, particularly as machine intelligence optimizes processes and feedback loops.
Do not forget the human touch
However, to get the most out of the next generation of business intelligence, automated data analytics must be combined with human intuition. Data sets, however refined, require human oversight for any resulting decisions to be fully informed and accurate. Increasingly modern platforms are being developed to "think like humans" using algorithms to analyze data from user interactions and make associated recommendations. In this vain, they will not only be generative, but also intuitive and associative in the same way that the human brain is – the ultimate cognitive machine.
These are characteristics linked closely with "cognitive computing" something distinct from solutions based on AI. The main difference between the two is that the AI used to power autonomous vehicles and consumer devices such as Amazon Alexa does not try to mimic human thoughts processes. Instead it learns from a specific set of variables and makes recommendations on the best possible algorithms for solving a pre-defined problem. For an autonomous car, for example, this would entail avoiding collisions and staying on course. In contrast, cognitive computing platforms extract contextual information as humans can, adapting as requirements and targets change. Unlike fixed algorithms they can resolve ambiguity and tolerate unpredictability, using probability to support decisions even with little representative data. Although such technology is still evolving and has a long way to go before mimicking the human brain, human attributes are being woven into analytics platforms themselves to support effective decision making.
Visit DATAx Singapore on March 5–6, 2019
Hurdling the bias barrier
With a combination of human intuition and a wealth of data at your fingertips, surely it is nearly impossible to make the wrong decision – but it does happen. Why do skilled data analysts, with all the information they have available, reach incorrect conclusions? Despite all the benefits of auto-generated data sets they should be treated with caution to avoid cognitive bias.
And this is arguably where cognitive computing really comes into its own. The defining characteristic of this type of platform is the provision of relevant factors and options for users to reach an ultimate decision, in contrast to a "black box" AI system that gives a recommendation that is based on a set of pre-defined algorithms. To counter the impact users should have access to data at a granular level to avoid only seeing part of "the big picture". Businesses need to be assisted by analytics to control bias and arrive at valid conclusions but the human ingredient is also key to shape the ultimate decision.
Making analytics accessible
Beyond the accuracy of the insights gleaned, augmented platforms enable business users get the most out of data without needing to understand algorithms or complex data models. After all, the uptake of analytics has until recently been hampered by the fact it requires human experts to find the right data, interpret results and come to the correct conclusion. Rather than replacing business intelligence tools or teams, augmenting users will expand adoption by helping them become more data literate and allowing them to uncover insights in an easier and more "governed" manner. Despite new solutions bringing humans, data and algorithms together, the former will always be required to interpret insights and make decisions.
Ultimately business is now being defined by the disruption of information, intelligence and advice. However, the cognitive era highlights the inextricable link between data analytics and human intuition, both in the design of generative platforms and the subsequent impact they have on the data literacy of business users. Despite the key role that automation and advanced algorithms must play in data analytics, the ideal model will always put humans at the center. This is augmented intelligence – the AI that is already shaping how businesses work.