Deezer will let listeners completely filter out AI music

Deezer plans to add an optional toggle by year-end that filters AI-generated music from listeners’ feeds, citing user demand. The company says AI-generated tracks have risen from about 18% of uploads last year to more than 50%, with 90,000 fully generated tracks uploaded daily and more than 42 million in total; Deezer has also reported that up to 85% of streams from AI-generated songs were fraudulent. The feature builds on Deezer’s AI-music detection tool, though the company says it is not foolproof and the risk of false positives is very low.

Analysis

The economic question is not whether listeners like a filter; it is whether platforms can convert trust into retention while reducing royalty leakage. If detected fraudulent streams are excluded from payout calculations, legitimate artists may receive a larger share of the pool, improving creator relations without requiring higher subscription prices. But a user-facing toggle alone does not prove that Deezer is removing those streams from settlement, nor that any savings accrue to the platform.

Deezer gains a plausible differentiation point, but the feature is unlikely to move consolidated economics unless it improves retention, acquisition, or creator supply. Spotify (SPOT) and Alphabet’s YouTube Music (GOOG) can respond with labeling or filtering; their scale and existing discovery funnels may limit Deezer’s advantage. A second-order risk for rivals is that visible AI controls raise user expectations and make weak detection or fraud enforcement more reputationally costly.

Near term, sentiment may favor DEEZR, but the product is not yet a demonstrated earnings catalyst. Over 1–3 months, watch for rollout timing, engagement/retention evidence, competitor response, and changes in detected-fraud stream treatment. Over 6–18 months, the structural issue is whether AI uploads degrade discovery and creator economics enough to shift users or artists between services. False positives could also suppress legitimate tracks and undermine the trust benefit. The thesis weakens if usage data show no retention lift, if competitors neutralize differentiation, or if detected fraud remains immaterial to royalty settlements.

Key Decisions for Investors

• No immediate trade: avoid extrapolating a feature announcement into material DEEZR earnings impact; verify rollout and any retention, subscriber-acquisition, or creator metrics first.
• Set an alert for Deezer disclosures on AI-detected streams excluded from royalty calculations. A measurable reduction in fraudulent payouts, without a rise in legitimate-track disputes, would strengthen the creator-economics thesis.
• Treat SPOT and GOOG as watch items rather than shorts: assess whether they add comparable opt-in controls and whether Deezer shows measurable share gains before positioning on competitive displacement.
• Falsification checks: delayed or limited rollout, evidence of meaningful false positives, no user-engagement improvement, or continued fraud dilution despite detection would undercut the differentiation thesis.