Smart Solutions for a Smarter Future: The Role of AI and Machine Learning
By RV Raghu
AI is one technology that is taking the world by storm, and rightly so. Over the years, wave after wave of technology has surrounded humanity and has brought with it change and disruption but none of these technologies has had as profound an impact as artificial intelligence (AI) and machine learning (ML). In some ways, the rise of AI and ML seem almost predestined, considering the data that is being generated and eventually used to feed their models.
As humans, we have always tried to envision a smarter, better future, one that takes us further, one that makes human lives better, leveraging technology to the greatest extent possible. This smart future is especially required considering factors such as an increasing life span, easy access to technology, democratization of information, the widespread availability of connectivity, automation in every sphere of life and the concomitant economic benefits from all this.
Globally there are pockets of excellence where the future seems to have already arrived reminiscent of what William Gibson once said about the future being here, but just not evenly distributed. Data and how it is used distinguishes these pockets of excellence and smartness from the rest of the world. The challenge lies in how this data is used, which can make it either a boon or a bane. The vast quantities of data can be difficult to manage for humans making AI and ML the next logical step.
Often, AI and ML can use data to learn about the world around us and make the world better. This works both at the macro and micro level. Data for example can be fed through AI models for better medical interventions so that custom drugs and physiological interventions can be tailored to an individual. There are also apps that leverage AI and ML to provide weather predictions with great precision and temporal proximity. Apart from this, of course, a combination of the cloud, edge computing, and AI and ML in personal devices such as smart watches and sensors is guiding us to live better lives than ever before.
At a macro level as well, AI and ML combinations are being used to manage traffic in cities, better route aircraft and cargo, and even manage other large-scale systems such as the internet and even national economies.
By their very nature, because AI systems can consume large quantities of data, understand patterns, and draw conclusions or ‘learnings’ from all this, they are able to make predictions about the real world which can be very useful. The fact that AI and ML systems can feed learnings from these predictions to update their algorithms and improve over time makes them powerful tools that can in turn make the world around us smarter.
Of course, all this is not a matter of serendipity and needs a close watch. For example, generative AI has taken the world by storm, and its large learning models (LLMs) at the center of this wave have become the darling of the media, enterprises, and the general public. A recent ISACA survey on generative AI indicated that cybersecurity and risk professionals in enterprises have concerns over the unbridled use of AI, with a mere 10 percent of respondents indicating their organizations have a formal, comprehensive policy for generative AI. The same survey also indicated that fewer than one third of organizations are prioritizing AI risks, which is worrisome. The challenge with using AI and ML tools is that the simplicity of the interface belies the underling complexities.
There are multiple areas that will need focus. First of course is the data—several problems can stem from the data used to train models including bias, ethical issues, and issues of fit. Second, enterprises will need to understand that from the outside, these models are black boxes; they may have no insight or control over the model or the algorithm, the data that was used to train it, weights, and conclusions that the model draws and how this can relate to the enterprise data that is fed in. Red flags have been raised on aspects such as explain-ability and applicability. Third, the absence of clear strategies and policies for the incorporation of AI in the enterprise is alarming and will need to be remedied. Fourth and probably most important is the rampant lack of skills within the enterprise at all levels about AI. This has an upstream/downstream effect in that decision making and incorporation of AI into the enterprise becomes very risky. In ISACA’s Next Decade of Tech study, 81 percent of respondents said enterprises are not investing enough in people skills needed to successfully navigate the changing landscape of the next decade. To be able to engender smarter solutions for a smarter future, enterprises should focus on building skills relating to AI and ML at an enterprise-wide level, starting with the board and the C-suite all the way down to front line employees, so that risks are effectively understood, and the benefits from AI can outweigh the risks.
(The author is RV Raghu, ISACA India Ambassador; Director, Versatilist Consulting India Pvt Ltd, and the views expressed in this article are his own)
