ScoreDetect details AI moderation rollout, 99% matching

ScoreDetect outlines a multimodal AI moderation stack with blockchain proof and a 90-day rollout plan.

ScoreDetect says AI can speed up content moderation by handling high-volume detection, clustering repeated violations, and preserving evidence for enforcement. The June 6, 2026 blog post also points to a 90-day rollout plan, a 99% matching engine, and blockchain-based verification for disputed cases.

What changed

The post frames moderation as a pipeline, not a single review queue. It breaks the workflow into signal collection, normalization, classification, triage, evidence packaging, and action, then maps AI tools to each stage.

On the detection side, the article says multimodal systems can catch altered or partial copies across images, video, audio, and text. It cites tools from InCyan such as Idem for matching and Tectus for blind watermarking, plus ScoreDetect for ownership proof.

• Idem is described as a multimodal matching engine with 99% accuracy.
• Tectus embeds invisible watermarks in the media signal, not metadata.
• ScoreDetect stores only an SHA-256 hash on the SKALE blockchain.
• The average blockchain transaction time is listed at 2.754 seconds.

Why it matters

For moderation teams, the pitch is less about replacing reviewers and more about reducing alert noise. The article says AI should cluster related sightings, cut reviewer fatigue, and move clear cases toward automated enforcement while humans handle edge cases.

For developers, the practical takeaway is that moderation systems need structured reference data, audit logs, and API-based enforcement hooks before AI can help. The blog’s 90-day plan starts with a charter, asset inventory, data-source mapping, and legal review, which makes the rollout as much an ops project as a model integration.

The bigger question is whether teams can build enough evidence quality and workflow discipline to make AI moderation defensible when takedowns are challenged.