Saregama MD Urges DSPs to Exclude AI-Generated Music

Saregama Managing Director Vikram Mehra told analysts during Saregama India's Q4 FY26 earnings call on May 14 that streaming platforms should not assign royalties to music "generated purely or through AI," as reported by Music Business Worldwide. Mehra said, "no value should get assigned to content which is getting generated purely or through AI," and added that AI-produced music is "there across all platforms" but "seeing no traction," with "no revenue leakage today," according to the same report. Mehra also stated that Saregama sees licensing opportunities with generative AI companies and said the company has launched a dedicated "AI efficiency team," per the earnings call coverage.

Companies arguing for exclusion of AI-generated content from royalty pools typically rely on two operational levers: transparent content labelling/metadata and automated detection systems that flag synthetic audio. Industry discussions involve tradeoffs between false positives in detection and the cost of human verification. Observers note that current audio-generation tools can imitate styles closely enough to complicate rights attribution unless a robust provenance or watermarking layer is adopted across platforms.

Industry context: Reporting captures a common 2026 tension between rights holders seeking to protect royalty pools and labels exploring new revenue via licensing to AI firms. Mehra's comments reflect that duality: a call to protect existing royalty distributions, and an acknowledgment that generative-AI licensing can be commercially interesting, as he noted on the call.

Key indicators for practitioners and platform engineers will include whether major DSPs update terms-of-service or royalty-allocation rules to explicitly address AI-generated audio; adoption of standardized content labels or watermarks by creators and platforms; announcements of licensing deals between major labels and generative-AI vendors; and advances in model-level provenance or detectability tools. Changes in these areas would affect catalog ingestion pipelines, metadata schemas, and rights-accounting processes.