Taxpayer Adaptation and Artificial Intelligence: Coretax Implementation Challenges

Fika Nur Sabrina (Universitas Terbuka, South Tangerang, Indonesia)

Jurnal Manajemen, Akuntansi, Ekonomi
Vol. 5 No. 2 (2026): Jurnal Manajemen, Akuntansi, Ekonomi (September)

This study examines taxpayer experiences with Coretax and the practical capabilities of mainstream AI platforms in performing Indonesian tax calculations. Through interviews with 26 taxpayers and experimental testing of ChatGPT, Claude, and DeepSeek, the research identifies significant gaps between reform objectives and on-the-ground reality. Findings indicate that 84.6% of respondents struggle with the Coretax system, with technical instability and insufficient outreach cited as primary barriers. Platform evaluation reveals that while Claude demonstrates the strongest regulatory alignment, DeepSeek applies outdated VAT rates, and ChatGPT produces calculation errors in complex scenarios. The study concludes that AI offers meaningful potential as a supplementary compliance tool, but its effectiveness depends on continuous regulatory updating, professional verification, and realistic expectations about capabilities.