Advancements in AI and Machine Learning

That the recent advances in Artificial Intelligence and Machine Learning is changing the landscape of job opportunities, is a gross understatement. Going by the recent announcements made by software giants, most of the code will be machine-generated by 2040, indicating fewer human hires in these organizations. The role of software professionals would be to innovate and design, rather than be stuck at the mundane job of writing code or testing.  While this does not definitely spell good for software job aspirants, the opportunities created by these advancements in other sectors is equally promising, if not more alluring for their novelty. From being technologies that were used only by expert computer scientists or data scientists, AI and Machine Learning technologies have now translated as “tools and weapons” necessary for every industry job, be it in the area of healthcare or finance or insurance or even content creation. All these are significantly reshaping job opportunities across industries, redefining the skills in demand and the nature of work, thereby creating entirely new career paths. 

The advent of agentic frameworks supported underneath by Large Language Models, is changing the way marketing teams operate. The agents are automating customer segmentation and campaign optimization, leaving the human experts free to work on insights, and strategies. Finance is also seeing an overwhelming adoption of AI to help in the analysis of complex, heterogeneous, multi-source data for improved fraud detection, risk modelling and algorithmic trading. The manufacturing domain is also seeing increased adoption of AI based predictive maintenance along with smart robotics co-occupying the work space along with humans. While all of these do mean fewer humans are needed for day to day operational jobs, the shift in trends also indicate a rising need for people who are trained to utilize the AI agents properly to their advantage, and not use them in a way that can backfire on the organization. Autonomous agents may sometimes lack real-world context due to unexpected situations that may creep into the workplace due to scenarios that were not envisaged properly, or due to circumstances beyond control. While deployment of AI agents can bring down operational costs significantly for a company, complete autonomy can lead to chaos, damage to life and property, and also unpleasant liabilities for the company that may affect business. These are precisely the reasons that are prompting companies to hire people who  are proficient in AI technology, tools and their applications, rather than domain specialists who are completely unaware about how AI works in their own domain. Natural language interface makes the interactions easier. Data-driven analysis is now less about navigating through the maze of data, but more about deriving insights, performing causal analysis and making informed choices about best alternatives to be adopted in future. Similar benefits are enjoyed by maintenance professionals in the manufacturing units. AI-driven preventive maintenance schedules are generated by AI-powered systems. In completely automated smart environments, these are done using the sensor data gathered from a multitude of sources. Both individual and collective analysis of equipment and resources promise to provide much more efficient workplaces in the near future.

Healthcare, finance, insurance and legal domains are some of the fastest adopters of advanced AI technologies. The kinds of complex data-driven decisions that these domains need can benefit greatly with the infusion of AI-powered automation. However, these domains also need decisions that are provable as ethically sound, fair and bias free, traceable and accountable. While the interpretation of bias-free and fair may vary from region to region, the very notion of fairness and bias hardly change. The world has not been bias free and fair. Thus there are numerous evidences on how AI systems trained on past data tend to amplify those biases, an act that is not acceptable by global standards. Another controversy that mires the use of AI models is the large volumes of copyrighted digital data that is used for training in many cases. 

A new set of organizational positions that are still in a very nascent stage, but promise to rise  steadily over the next few years are those of professionals whose task is to ensure practice of responsible AI  that align with human values and societal well-being. These positions are not confined to computational scientists or domain experts. Rather, these constitute a diverse array of professionals whose task is to assess the AI systems from different perspectives right from design to deployment. While at deployment stage , it is more about ensuring fairness of decisions for different groups of society, the scope is much wider which also oversees ethical uses of data both at collection points, and during model building. Linguists, behavioural specialists, social scientists, culture exponents all have a role to play in this. As AI gears up to reach far and wide, demand for certified Ethics professionals will be on the rise. 

There is also a rise in the trend of social entrepreneurship designing innovative applications aimed at solving problems for people and the planet.  AI technologies like drones, language and vision technologies, smart sensing and robots are revolutionizing sectors like agriculture, healthcare, insurance and education. Starting with data-driven analysis to identify the root causes for problems like water scarcity or health hazards or social inequity, the entrepreneurs are looking at solutions that can bring about a 360 degree change to the regions.  Deep technical knowledge along with a zeal to solve social problems complemented by people skills is the ideal winning combination for success in this area. 

While the recent trends in AI and ML have brought about a total transformation in the way technology-driven solutions are built for a wide range of problems, there is also a rising concern about the energy use by these technologies. The power of generative AI comes at the cost of using energy that is enough to sustain hundreds of households for one full year.  With the proliferation of data centres to support this growth, the impact on the earth is significant. Alternatives focusing on low energy substitutes of smart technology should be also encouraged. Building Edge AI platforms and smaller generative models are some of the areas that the research community is focused on. Hopefully, these initiatives will also gain commercial footholds soon. 

Author: Lipika Dey, Professor, Computer Science, Ashoka University