Both stats come from a BPI-commissioned WPI Economics report based on surveys of labels and consumers. Technically published last month, said report was cited by the Featured Artists Coalition today in a release concerning 29 organizations’ consent-focused open letter about AI dealmaking.

As for the little-discussed scope of across-the-board AI pacts, 274 such “commercial agreements” had materialized “from a range of creative sectors” as of early 2026, according to the analysis.

Regarding the absence of the all-important “generative” descriptor, a few of the tie-ups center on content fingerprinting and other adjacent areas. In music, this refers to arrangements between Sony Music and Vermillio as well as Universal Music and Prorata.ai, to name a couple.

Nevertheless, the lion’s share of the industry’s AI deals do, of course, involve gen AI. Just in passing, the report recaps related unions with companies including Suno, Udio, Klay Vision, Spotify (all three majors, Believe, and Merlin), ElevenLabs (Kobalt, Merlin, and SourceAudio), BandLab, and Splice.

Additionally, the AFM is suing for a piece of the relevant majors’ Suno and Udio licensing revenue. And the above-noted 29 organizations are criticizing labels’ and publishers’ various alleged efforts to compel AI opt-ins without painting a complete picture (in terms of revenue and otherwise) for creatives.

Put differently, mountain of music-space AI agreements aside, the sub-sector certainly isn’t without controversy and disputes. That leads to another interesting element of the report, for which a survey was forwarded “to a wide range of independent BPI member organisations.”

Of those members, 31 responded during a 16-day window, and eight proceeded to participate in “semi-structured interviews.”

At the intersection of both figures, “only 16% of independent BPI members surveyed have so far begun exploring licensing partnerships” at all, the breakdown shows. Even so, 77% of member respondents “are open to licensing their music for ethical use by AI.”

Beyond this noteworthy gap, the report produced an interesting takeaway about the public’s stance on rightsholder compensation for AI music.

Of the overall paid-AI respondents, 65% agreed that any “financial gains to the economy from AI should be shared beyond the tech sector, including with those whose work is used in AI training.”

Meanwhile, the percentage fell to 60% among free-tier AI users and then to 47% for respondents who said they’d never used AI.