In a striking revelation, AI technology played a pivotal role in identifying GST mismatches amounting to a staggering ₹30,000 crore during the previous financial year. Senior officials at the finance ministry disclosed that they have successfully recovered ₹15,000 crore of the identified amount, underscoring the significant impact of artificial intelligence in tax administration.
A substantial surge in scrutiny has been witnessed, with central tax officers dispatching approximately 33,000 GST notices to businesses concerning discrepancies in returns for the 2017-18 and 2018-19 periods. This sharp increase in notices has sparked apprehensions among businesses, with concerns amplified by a senior government official confirming AI’s primary responsibility for detecting tax evasions and consequently issuing a higher number of notices.
Specific industries have come under intense scrutiny, including online gaming companies, which faced 71 show-cause notices for alleged GST evasions amounting to a staggering ₹1.12 lakh crore between 2022-23 and the initial seven months of 2023-24. Notably, 15 major insurance companies, including prominent names like Bajaj Allianz, Aditya Birla Sun Life Insurance, and HDFC Life Insurance, are accused of evading ₹2,350 crore in GST payments. Further, sectors such as aviation and shipping have also piqued the interest of tax authorities.
While acknowledging AI’s effectiveness in detecting trends, experts caution against sole reliance on this technology for issuing notices. One expert stressed the importance of human intellect in tax administration, stating, ‘AI is a powerful tool to identify trends, but its blind utilisation by the government to issue notices could lead to legal implications.’
Contrary to assertions, officials from the finance ministry’s revenue department clarified that AI is utilised solely for identifying payment discrepancies, after which human intervention determines whether a demand notice is warranted for a particular case. However, concerns persist regarding the prevalence of notices solely triggered by AI without substantial human assessment.
Another expert emphasised that while AI expedites data review and gap identification, it should not serve as the sole basis for issuing notices. The need for tax authorities to engage in discussions with taxpayers regarding factual and legal aspects before issuing notices was stressed for effective utilisation of AI and emerging technologies.
Addressing the need for improved communication between industries and the government amid increased AI reliance, experts advocated for issuing intimation notices before show-cause notices. These notices, a provision within the GST law, inform taxpayers about discrepancies and seek explanations within 30 days, reducing the element of surprise and friction.
Reflecting on the evolving landscape of tax administration, officials noted the increasing sophistication brought about by AI and standardised legal interpretations. They highlighted the diminishing areas where businesses could conceal discrepancies owing to the evolving case law and standardised data analytics.
The growing reliance on AI in uncovering tax evasions signals a paradigm shift in tax administration, necessitating a balanced approach that harnesses technology while ensuring human oversight to maintain legality and fairness in the process.
With over a decade of experience in the legal journalism field, Rajeev has a knack for simplifying complex legal matters for the common reader. His dedication to unbiased reporting and commitment to factual accuracy has established him as a respected figure in the Indian legal community. When he's not dissecting the latest court rulings, Pravien enjoys mentoring young journalists.
- India’s Economic Pulse Beats Strong as LLP Registrations Soar 54% - January 5, 2024
- Company Faces ₹5 Lakh Penalty for Breaching Registered Office Compliance: Ministry of Corporate Affairs Takes Action - December 13, 2023
- AI’s Growing Role in Uncovering GST Evasions Raises Concerns Among Businesses - December 13, 2023
