Impact on Skills Development

Impact on Skills Development

There has been tremendous progress conceptualised and delivered in the last few years in the area of skills. A case in point is corporate learning, where organisations have procured hundreds of e-learning course modules for their employees only to find that both usage and course completion rates have been abysmal. The reasons for this poor response are many but can mainly be attributed to a feeling of ‘why on earth should I learn this stuff?’ in employees and the inability of purely e-learning approaches to personalise the learning to individual needs.

Enter AI, and the entire approach and pedagogy can be transformed. A wonderful India-developed platform, Skills Alpha, works through a five-phase model that is transforming learning effectiveness in organisations by:

1. Capturing previous appraisal and performance data about the employee
2. Engaging through a bot to discuss the career aspirations of the person and explore various options and the unlearning, reinforcement learning and new learning needs to be qualified for the chosen role
3. Analysing the track record of learning that the person has demonstrated through available data and choosing the right pedagogy for maximum learning effectiveness. This could be as simple as choosing video courses rather than text or increasing levels of complexity in course design for the individual.
4. Measuring learning effectiveness through the courses completed by the person and also integrating the outcomes of courses taken outside the learning platform with quantitative and qualitative data capture
5. Assigning a mentor who has traversed the track a few months or years earlier to guide the person over learning doubts and difficulties and later help the person to navigate situations that may have an adverse career impact

The future of AI in enabling better skills development is only limited by our imagination. With so much data available as the social construct of each individual on the internet, it is within the realm of reason to assume that every step of skills development can be re-engineered for learning effectiveness. Beyond the skilling process itself, there are many contemporary skills that get individuals jobs and careers that can be transformed by AI. Cases in point are many and can be broadly categorised in three buckets.

1. Industry-related skills, where awareness needs to be built on the nuances of the industry that people are getting trained for. In the previous chapter, we have detailed the opportunities for AI in many industry sectors, and job seekers in these sectors would do well to be trained by AI-on-AI opportunities. Whether it is superior shop floor techniques for the manufacturing industry or intelligent predictive analytics for people entering healthcare, every industry has nuances that need to be quickly learnt and mastered to build career success.
2. Technical skills can also be taught with large doses of AI input. A welder can master his craft better if he is watched and guided by AI on better techniques suited to his hand flexibility, and a supervisor managing an automated plant can be guided on potential failure points to look out for through years of past performance and failures and likely future occurrences.
3. Soft skills are where AI can probably play a very substantial role in building better supervisors and managers. Every situation is different in a working environment, but there have been situations where decisions taken on impulse without adequate data have resulted in suboptimal outcomes. Predictive and prescriptive analytics can be brought into play to enhance the percentage of good decisions made by managers and leaders. AI can play a significant role in enhancing the style flexibility of supervisors too. AI can analyse the behaviour patterns of each member of a team and guide the supervisor on responses as well as the mode of taking corrective action with any delinquent subordinate. Over thirty years ago, one of the authors had an opportunity to go through an interactive learning course where an aspiring leader could sit at a console with seven or eight people at a conference table on the video screen in front of him or her. In a situation where one member seemed to be dozing off in the meeting, the leader would be presented with three or four options: to ignore, to request attention or to admonish in a loud voice. Watching the response would guide towards a better action in each situation. Today, with AI researching every past action and outcome for each individual, the guidance could be very granular, customised to each team member. The possibilities are truly infinite, and much can be predicted and used to guide actions in future.

The most endangered skills will be those that have enabled nearly a million Indians to build global careers in the world of IT and BPO services for third-party as well as global capability centres that dot the Indian landscape. These are programming or coding, testing and offshore maintenance services. Coding has for long been the sought-after skill for youth across the country, and it is necessary to understand whether this will become entirely irrelevant in the years to come or whether there are still areas that young career aspirants can and should pursue.

AI is certainly capable of generating code for any development environment and can also correct or complete code written by programmers for previous applications. Detecting and fixing bugs, testing for completeness and optimisation are all within the purview of AI code generators. Development cycles in large systems development projects can definitely be accelerated by applying ‘dual intelligence’, an approach that has been advocated by new data company GTT Data Solutions as the optimal approach to combine human and AI to tasks and programs of the future.

What then are the areas where human intelligence will continue to score over AI, at least in the foreseeable future? When one of us asked our personal AI what it saw as its current limitations, it quickly confessed to falling short in human judgement and empathy, context and intent understanding, creativity and navigation of ambiguity and complex decisions.

The passionate trainers of LLMs would scoff at these and throw a couple of graphics processing units at the AI to overcome these limitations faster. The Chinese creators of DeepSeek would enhance its distillation and inference capability to discern emotions and modify behaviour based on what it perceives. However, there is no doubt that the data-information-knowledge-wisdom quartet that many of us have traversed will take AI a lot of learning to master. This may raise the issue of what a young inexperienced software engineer should learn to start his career in these complex times, which may have to be answered differently with every cycle of AI evolution.