By Anurag Srivastava
With generative AI becoming paramount to businesses, there has been a significant rise in demand for artificial intelligence (AI) skills. A major shift in the division of labour between machines and humans is afoot with the emergence of 97 million new jobs by 2025 owing to the sudden demand for AI skills such as prompt engineering and problem formulation. While prompt engineering skills have and continue to get the spotlight, the equally crucial skill of problem formulation is going under the radar.
Problem formulation is the precise definition of an issue to make it suitable for AI methodologies. This process transforms a real-world problem into a computational challenge that algorithms can address effectively. Not only does problem formulation form the basis of prompt creation, but it also requires a more holistic approach of the problem’s real-world intricacies, which will ensure that this skill stays in demand in the future and avoids obsoletion by AI itself.
This skill will prove to be central to harness the potential of GenAI, and enhance the problem-solving process in these ways:
Provide clarity for the problem at hand by incorporating domain expertise and outlining desired outputs, biases, and constraints
Make the problem-solving process more efficient by outlining the scope of the problem, breaking it down into manageable subproblems and prioritising amongst these, which enables better resource allocation and time management
Ensure that the solution is aligned with the user’s requirements and desired outcomes
Help in mitigating biases and unethical behaviour as it allows for the establishment of constraints and safeguards that guide the AI system’s behaviour
Due to its iterative nature, problem formulation results in continuous reinforcement for AI models. This ongoing process fosters gradual enhancements in the models’ natural language processing (NLP) capabilities, ultimately bolstering their effectiveness from a long-term standpoint. This skill can be further honed and enhanced by:
Acquiring a good understanding of the problem domain
Developing one’s critical thinking skills
Learning from existing solutions available in the public domain.
An interesting aspect is the degree of human intervention required across its touchpoints and throughout its stages. Given the intricate and highly contextual nature of human language systems, the iterative nature of this process means that this human intervention results in the gradual refinement of the AI solutions’ computational capabilities.
The capability to accurately define and articulate the problem becomes of utmost significance in the age of generative AI. By grasping problem formulation, individuals can navigate the complexities of generative AI, effectively harness its potential, and mitigate potential risks or unintended consequences.
The writer is senior vice-president at Findability Sciences, an enterprise AI company