U.S. consumers are tilted toward caution rather than acceptance when generative AI enters content creation. That is the finding of AI & Media: Audience Attitudes, the special report Luminate Intelligence (루미네이트 인텔리전스) published on July 24. The spread between film, TV, music, social media and video games was narrow.
Jung Han
“more wary than willing”
— Luminate Intelligence, AI & Media: Audience Attitudes (July 2026)
Three shifts landed in the same period, which is what gives the survey its weight. Generative AI moved from technical steps such as subtitling, translation and post-production into creative ones such as scripting, vocals and character design; Spotify and Apple Music each switched on track-level AI labeling this year, creating a real point at which listeners learn AI was involved; and Korea brought its AI Framework Act (인공지능 발전과 신뢰 기반 조성 등에 관한 기본법) into force on January 22, making disclosure of AI-generated output a legal duty. The question of how audiences react once they know has moved from hypothetical to operating condition.
Age split the audience more than media type did
The report frames three variables behind willingness to engage: what material AI is generating and what purpose the content serves (media type); how extensively AI was used in the creative or production process, and whether the output is entirely or only partly AI-generated (extent of use); and whether AI has taken over a process or output whose creative and communicative intent belongs to a human (creative versus technical process).
Differences by media type were narrower than expected. The strongest opposition attached to reading news written with gen AI; the weakest attached to playing a video game containing AI assets. Gaming's relative standing came from indifference rather than enthusiasm: 40% said they would be neither more nor less interested in a game whose characters, environments or dialogue were made with AI, against roughly a third ambivalent toward AI uses specified in other media.
[Fig. 1] Neutral responses by media type (Source: Luminate U.S. Entertainment 365, Wave 17)
The age cut is where the gap appears. Luminate broke the data out by age band rather than by generation for the first time in this wave. Consumers 55 and older registered the deepest net-negative readings across film, TV, music, social media and gaming. At the other end, the 13-17 cohort was net positive across every media type, and the distance to the 18-24 cohort reached 58 points on both AI-scripted movies and TV and AI-generated social media content. Older Gen Z, Millennials and Gen X clustered near breakeven.
[Fig. 2] Where each age band sits on net interest in AI-assisted content (Source: Luminate, May 2026)
“it would be a mistake to regard Gen Z as a monolith”
— Luminate, as reported by Music Business Worldwide
Luminate added that intragenerational faultlines are the more useful lens.
News draws the thickest line of resistance
News written with gen AI drew the strongest opposition in the survey. Set against news consumption data, the two trends run in different directions. The Digital News Report 2026, published by the Reuters Institute for the Study of Journalism (로이터저널리즘연구소) on June 16 and based on roughly 100,000 respondents across 48 markets, recorded weekly use of AI chatbots for news rising from 7% in 2025 to 10% in 2026. Among under-35s the figure was 16%, while only 1% named AI as their main source of news.
[Fig. 3] Where AI chatbots sit in news consumption (Source: Reuters Institute, Digital News Report 2026)
The same study put trust in news at 37% and news avoidance at 42%, with social media and video networks reaching 54% to become the most widely used news channel for the first time. Audiences are increasingly reaching news through AI while pushing back on news written by it — a distinction between AI as a tool and AI as an author.
“the first meaningful steps towards the integration of AI into news journeys”
— Reuters Institute, Digital News Report 2026, executive summary
Music: the line falls between assisting and replacing
The item-level readings suggest consumers are judging AI's role rather than AI in general. One in three U.S. music listeners said they were comfortable with AI producing the instrumental for a song performed by a human artist. By contrast, 46% were uncomfortable with an entirely new song performed by an AI voice.
[Fig. 4] Consumer response by type of AI use in music (Source: Luminate reports)
“Acceptance turns on whether the technology supports human creation or takes its place.”
Generational temperature sits on the same axis. Luminate's 2025 Year-End Music Report found about 44% of U.S. respondents would be less interested in music once they knew generative AI had produced it, while 37% of teenagers and 36% of Millennials said they had no problem with AI-generated lyrics. Over time the line points down: in Luminate's Generative AI in Music tracking, net interest in AI-generated music fell from -13% in May 2025 to -20% by the end of that year, with the steepest decline among Gen Z and Gen Alpha.
[Fig. 5] Net interest in AI-generated music over time (Source: Luminate, Generative AI in Music)
There is also academic evidence that the label itself moves demand. A peer-reviewed study by researchers at Stanford University and Reed College published this year found listeners engaged less deeply with music marked as AI even when a human had composed it. Disclosure requirements, on that reading, reach usage metrics directly rather than stopping at compliance.
The labeling infrastructure arrived first
Platform plumbing was largely built in the first half of this year. Spotify opened its AI Credits beta on April 16, surfacing DDEX-standard AI-use declarations inside song credits, with DistroKid as the first distribution partner. Apple Music introduced Transparency Tags in March across four categories — artwork, track audio, composition and music video. Both systems leave the judgment of what counts as material AI use to labels and distributors, making disclosure self-reported. Spotify said it had removed more than 75 million spam tracks in the year before its September 2025 policy update.
Deezer sits on the detection side. It automatically identifies and tags fully AI-generated tracks and keeps them out of editorial playlists and algorithmic recommendations. Roughly 90,000 such tracks arrive daily, crossing half of the platform's daily uploads for the first time in June.
[Fig. 6] Fully AI-generated tracks as a share of Deezer's daily uploads (Source: Deezer, June 2026)
From litigation to licensing
The rights picture also turned within a year. Universal Music Group settled with Udio in October 2025, and Warner Music Group settled with both Udio and Suno in November, pairing the settlements with licensing agreements. Sony Music has continued to litigate. Udio's 2026 model moves to a walled-garden design in which creations do not leave the platform.
“artists will have full control over whether and how their voices are used”
— Warner Music and Suno joint statement, via Music Business Worldwide
The settlements themselves have now drawn a suit. The American Federation of Musicians alleges that Universal and Warner licensed member session recordings to Suno and Udio without the compensation and credit their collective agreement requires. Labeling, licensing and compensation are being settled on three separate layers at once, and consumer data applies a different kind of pressure on each.
What this means for Korean players
Three operational questions follow for Korean companies (한국 사업자).
The first is how AI use is disclosed. Article 31 of the AI Framework Act requires operators to notify users in advance that a product or service runs on generative AI and to mark output as AI-generated.
“shall notify users in advance that the product or service operates on that AI”
— AI Framework Act, Article 31(1), author's translation
For ordinary output, businesses may choose between a human-perceptible notice and machine-readable methods such as watermarks or metadata; for deepfake output, only a clearly visible notice is permitted. With a grace period of at least a year, fines of up to KRW 30 million are unlikely before 2027. Election law is already at the enforcement stage: deepfakes made or distributed for campaign purposes are banned from 90 days before an election, AI-generated material must be labeled outside that window, and the National Election Commission (중앙선거관리위원회) began enforcement on March 5 ahead of the 2026 local elections, producing its first referral. Whether a label reads simply “made with AI” or specifies something closer to “AI-assisted post-production” maps directly onto the acceptance gaps this survey documents.
“the effect on creative freedom, competitiveness and user experience also needs thought”
— KOCCA (한국콘텐츠진흥원) TrendOTT, on AI-output labeling, author's translation
The second is genre allocation. News and current affairs drew the strongest resistance to gen AI involvement, while gaming and interactive formats carry the thickest neutral bloc. As the Reuters data shows, audiences are expanding their use of AI to reach news while rejecting it as author. For news and documentary work, that argues for keeping AI in the research, search and translation layers and out of the final text; for gaming, virtual and interactive formats, it argues for putting AI use in the foreground.
The third is targeting. A 58-point gap between 13-17s and 18-24s indicates that “Gen Z” is too coarse a unit for campaign design. In audiences with wide age spread, such as K-pop (케이팝) fandoms, AI-related messaging has to be scored on the attention it gains in one segment against the trust it costs in another. Replacing an artist with an AI voice and using AI for instrumentals or audio post-production both file under “AI use,” yet the market responses run in opposite directions. Warner and Universal building artist opt-in into their deals is aimed at the same seam.
Methodology
Luminate's U.S. Entertainment 365 is an original online survey fielded among U.S. respondents aged 13 and older, representative of the general population per U.S. Census for age, gender, ethnicity and region. This report draws on Wave 17, fielded May 5-26, 2026, among 2,000 respondents. Luminate notes that methodological updates applied in Wave 17 affecting demographic proportionality may cause slight deviations from its previously published figures. The Digital News Report 2026 surveyed roughly 100,000 online news users across 48 markets.
Fig. 2 plots only the direction and the gap disclosed in the report; per-age-band values were not published. Fig. 4 gathers items from different questions and waves, so it should be read directionally rather than as a like-for-like comparison.
Sources
1. Luminate Intelligence, AI & Media: Audience Attitudes — Examining Consumer Acceptance of Generative AI in Music, Film and TV, Special Report, July 2026 (source PDF)
2. Audrey Schomer, “How U.S. Audiences Feel About Gen AI in Entertainment Content,” Luminate Spotlight blog, July 24, 2026 — https://luminatedata.com/blog/how-u-s-audiences-feel-about-gen-ai-in-entertainment-content/
3. Music Business Worldwide, “US consumers are ‘more wary than willing’ to engage with AI in entertainment content, Luminate study finds,” July 2026
4. Music Business Worldwide, “The Music Industry Is Trying to Guess the AI Recipe After Dinner Was Served,” August 2026 (music figures and Deezer upload data)
5. Luminate, 2025 Year-End Music Report and Generative AI in Music — net interest trend (-13% to -20%) and generational breakdowns
6. Reuters Institute for the Study of Journalism, Digital News Report 2026, June 16, 2026 — https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026/dnr-executive-summary
7. Music Business Worldwide, “Warner Music Group strikes ‘landmark’ deal with Suno; settles copyright lawsuit against AI music generator,” November 2025
8. Billboard, “What Do the Suno and Udio Licensing Deals Mean for the Future of AI Music?,” March 2026 (Udio's walled-garden shift)
9. Forbes, Virginie Berger, “Launch, Train, Settle: How Suno And Udio's Licensing Deals Made Copyright Infringement Profitable,” December 2025 (settlement timeline and opt-in structure)
10. Korea Creative Content Agency (한국콘텐츠진흥원, KOCCA), TrendOTT, on AI-output labeling obligations — https://www.kocca.kr/trendott/vol02/spotlight_4.html
11. AI Framework Act (인공지능 발전과 신뢰 기반 조성 등에 관한 기본법, Act No. 20676), Article 31 and its Enforcement Decree, effective January 22, 2026 — Korea Law Information Center (국가법령정보센터)
12. Shin & Kim (법무법인 세종·신앤김) newsletter, “The AI Framework Act and the Content Industry — Focusing on Transparency Obligations,” February 2026 — https://www.shinkim.com/kor/media/newsletter/3142