Perkembangan era digital mendorong artificial intelligence (AI) menjadi salah satu pencapaian teknologi terpenting sepanjang sejarah, tetapi kehadirannya juga menimbulkan isu terkait transparansi algoritma atau yang dikenal dengan istilah “black box AI”. Ketidaktransparanan algoritma ini membuka peluang terjadinya pelanggaran hukum, termasuk di bidang persaingan usaha, di mana pelaku usaha dapat menyisipkan mekanisme yang menguntungkan mereka tetapi merugikan konsumen atau pelaku usaha lain. Praktik seperti predatory pricing, price fixing, atau keberlangsungan kartel menjadi beberapa contoh risiko yang dapat terjadi akibat penyalahgunaan AI. Penelitian ini bertujuan untuk mengidentifikasi permasalahan dan menawarkan solusi terhadap isu tersebut dengan penelitian hukum doktrinal yang bersifat preskriptif, menggunakan data sekunder sebagai sumber utama. Analisis data dilakukan dengan model Miles dan Huberman yang meliputi reduksi data, penyajian data, dan penarikan kesimpulan. Hasil penelitian menunjukkan dua solusi potensial, yaitu pembentukan AI database dan penerapan AI auditing sebagai bukti tidak langsung (indirect evidence), yang memerlukan kolaborasi intensif antara para pemangku kepentingan, termasuk Komisi Pengawas Persaingan Usaha (KPPU). Melalui solusi ini, permasalahan black box AI diharapkan dapat diatasi, sehingga transparansi dalam penggunaan teknologi AI di sektor persaingan usaha dapat meningkat.

---

The rapid development of the digital era has positioned artificial intelligence (AI) as one of the most significant technological achievements in history. However, its emergence also raises issues concerning algorithmic transparency, commonly referred to as “black box AI”. The lack of transparency in algorithms creates opportunities for legal violations, particularly in the field of competition law, where businesses may exploit AI to embed mechanisms that benefit themselves while harming consumers or other businesses. Practices such as predatory pricing, price fixing, or sustaining cartels represent some of the risks associated with AI misuse. This study aims to identify the problems and propose solutions to address these issues through a doctrinal legal research with a prescriptive nature, utilizing secondary data as the primary source. Data analysis was conducted using the Miles and Huberman model, which includes data reduction, data display, and conclusion drawing. The findings reveal two potential solutions: the establishment of an AI database and the implementation of AI auditing as indirect evidence, which require intensive collaboration among stakeholders, including the Indonesian Competition Commission (KPPU). These solutions are expected to address the issue of black box AI, thereby enhancing transparency in the use of AI technology within the competition law sector.