ArA-DF-2026: Arabic Speech Deepfake Detection Audio Dataset

ArA-DF-2026 is an Arabic speech deepfake detection dataset for binary audio classification, distinguishing between spoofed/synthetic and bona fide speech. The dataset includes labeled train and development splits, with unlabeled development-test and final-test splits scored externally via CodaBench. It was created by ArabicSpeech and last updated on June 16, 2026.

Use Cases

• Train binary classifiers for speech deepfake detection based on the labeled audio data.
• Benchmark model performance on unlabeled test splits via the external CodaBench platform.
• Research the characteristics of synthetic versus bona fide Arabic speech signals.

Strengths

• Audio is provided in a consistent 16 kHz mono PCM format packaged as lossless FLAC files.
• Includes a clear binary classification scheme (0 for spoofed, 1 for bona fide).
• Dataset was last updated on 2026-06-16, suggesting recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and total dataset size are unknown, which may limit suitability assessment.
• The description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

ArabicSpeech

Freshness

Last updated 2026-06-16 06:26:27.