Article Preview
Top1. Introduction
Technology interventions like digitalization and Machine Learning (ML) have had a commendable impact on public services (Cheng, Hu, & Wu, 2021) and consumer behavior (Ahmad, Masri, Chong, Fauzi, & Idris, 2020). Technology applications not only help to enhance public services but also reduce deviant consumer behavior (DCB). (Fullerton & Punj, 1997) defines deviant consumer behavior as any behavior which is “against the law, organizational policy or violates the generally accepted norms of conduct.” DCB causes financial and physical losses to the organization and emotional harm to the owners and employees (Daunt & Harris, 2012).
Organizations, especially public service organizations, use tactics like communicating with customers to comply with the legal and social norms centering their messaging around, “it’s wrong, don’t do it.” The second tactic that industries such as the retail industry use are evoking fear of punishment. In these tactics, organizations have to proactively demonstrate that customers cannot get away with unethical practices and that they may be caught and punished for their deviant behavior. The second tactic is defined as the “Deterrence tactic: you will be caught and punished” (Dootson, Lings, Beatson, & Johnston, 2017).
Organizations use hardware like CCTV cameras and non-hardware solutions like analytics, Artificial Intelligence (AI), and ML to capture deterrence tactics. Analytics is used to increase enterprise value by appropriate application in several functional areas, viz data to increase sales, and improve customer service and operations, to name a few (Baker, Al-Gahtani, & Hubona, 2010). It is also finding extensive use in other activities ranging from predicting train tickets and confirmations to checking for water supply leaks and even finding the perfect bride and groom. Governmental services are using analytics to combat crime, improve transparency, and services such as transport, etc. (Raghupathi & Raghupathi, 2014).
Increasingly, energy utility firms are using analytics to optimize power generation and planning (Kim et al., 2016). However, energy theft remains their major concern, adversely affecting the bottom-line and profitability (Dick, 1995). Electricity losses in utility firms are recorded under two heads, namely Technical Losses (TL) and Non-Technical Losses (NTL). Power dissipation in transportation and distribution of power falls under TL. Commercial losses are due to non-billed electricity, defined as non-natural losses and recorded under NTL. Non-billing of consumed electricity happens due to errors in metering or non-legitimate behavior of consumers (Oliveira et al., 2001).NTL reduces the finance available with utility firms for investing in further growth (de Souza Savian et al., 2021).
Emerging economies face the brunt of energy thefts; for instance, Brazil and India record an annual loss of USD 3 billion (Z. Hussain, Memon, Shah, Bhutto, & Aljawarneh, 2016). Extant research has adopted various methods, including AI-based, game theory-based, and state-based models, to capture NTL. ML and deep learning (DL) are constituents of AI-based approaches. ML is the process of training a machine with an algorithm to handle large data efficiently by predictive analysis. On the other hand, DL (Mohammad, Thabtah, & McCluskey, 2012) is based on an artificial neural network (ANN), a human brain model that helps to model irrational functions.
Using AI, utility firms can detect usage patterns, payment history, and other customer information that indicates misconduct (Gunturi & Sarkar, 2021; J. Li & Wang, 2020). For instance, in Brazil, power theft represents up to 40 percent of the distribution of electricity, while India loses approximately 25% of its supply, amounting to INR 200 billion every year (Gunturi & Sarkar, 2021). Hence, the Indian energy utility space calls for an urgent application of AI (Akter et al., 2021) and ML to capture and address deviant consumer behavior.