Anti-AI sentiment attaches to visible AI: labels, generated media, chatbots and data centers

AI becomes invisible by 2030: a background layer that takes care of things and is taken for granted by most, woven into infrastructure, devices, work and family life, with fewer labels, fewer exposure points and fewer visible surfaces.

The sentiment research of 2025 and 2026 locates the pressure on the surfaces people can see. In 16 experiments with 27,491 participants, the same creative writing was valued lower once labeled as AI-generated. In listening tests, participants rated AI-generated songs above human-made ones until told the origin, after which willingness to pay fell. Adding the word "AI" to a product description lowered purchase intention in every category tested. 71% of Americans oppose an AI data center in their own area, above the 53% who oppose a nuclear plant. 52% of US adults are more concerned than excited about AI, up from 37% in 2021, and 55% of adults under 30 said the same in June 2026, the first majority in that age group. Spam filtering, fraud scoring, camera autofocus and route planning have run on machine learning for years without a label, and none of the surveys cited below asks about them.

The label lowers the rating

The experiments below hold the content constant and change only what participants are told about its origin.

Survey data on brand content runs the same way. Capgemini tracked trust in AI-generated content from 73% in 2023 to 55% in 2025, with declines in every age group including Gen Z. YouGov's 2026 data has 32% of consumers trusting a brand less for using AI-generated content and 15% trusting it more. A 2026 Gartner survey has 50% of US consumers preferring brands that keep generative AI out of customer-facing messages, ads and content.

Where the hostility sits

The measurements below cover generated content, chatbots, hiring screens, employer rollouts and buildings.

Merriam-Webster chose "slop" as its 2025 Word of the Year on December 15, 2025, defined as low-quality digital content produced in quantity by AI. Macquarie chose "AI slop", Collins "vibe coding" and Oxford "rage bait" in the same season.

Pew, Bentley-Gallup and Gallup use probability samples. Luminate, YouGov, Writer and Capgemini use online opt-in panels, and Luminate, Writer and Gartner are vendor studies. The KPMG/Melbourne study is academically led and co-funded by KPMG.

The data center is the physical surface

Gallup asked about local data centers for the first time in March 2026 (1,000 US adults by phone, March 2-18, margin of error 4 points). 71% opposed building an AI data center in their area, 48% strongly, against 7% strongly in favor and about a quarter in favor overall. The same poll found 53% opposed to a local nuclear power plant. In a follow-up open-ended question, half of opponents cited resource use, with 18% naming water and 18% naming energy, and 16% cited pollution including noise. 46% said they worry a great deal about the environmental impact of data centers. An Annenberg Public Policy Center poll put opposition to new local data centers at 61%, with 69% of Democrats and 54% of Republicans opposed.

Data Center Watch, a tracking project run by the AI-security firm 10a Labs, counted $64 billion of projects blocked or delayed by local opposition between May 2024 and March 2025, $156 billion delayed or cancelled across 2025, and 75 projects worth about $130 billion in the first quarter of 2026 alone, the largest quarterly figure since its tracking began in 2023. Its count of active opposition groups doubled, from 396 to 833, across 49 states. The project's own report states that attributing delays to activism alone oversimplifies the causes. Estimates of per-query water and energy use vary widely between sources.

The issue entered the 2026 midterm campaigns. A National Republican Senatorial Committee memo reported by Axios warned that data centers were hurting the party in an Ohio race. On a primary day in August 2026, Pennsylvania Governor Josh Shapiro signed an executive order setting stricter standards for data center development, and Byron Donalds won Florida's Republican gubernatorial primary while proposing restrictions on data centers in the state. Two beverage brands, Liquid Death and Garage Beer, released a satirical ad with Jason Kelce about data center water use on the same day. On June 25, 2026, Mark Cuban wrote on X that data centers had become "a proxy for the hate towards AI"; Fortune reported the post drew more than 700,000 views within hours.

Adults under 30 crossed 50% in 2026

In 2021, adults under 30 were the most optimistic group Pew tracked: 25% were more excited than concerned and 31% more concerned than excited. In June 2026 the figures were 11% and 55%, and 73% of under-30s expected AI to reduce jobs over the next 20 years, up from 61% in 2024. Concern among adults 30 and over rose mainly between 2021 and 2023 and then flattened, while concern among under-30s kept rising through 2025 and 2026 until it matched people in their 30s, 40s and 50s. The under-30 subsample was 569 people with a margin of error of 4.7 points.

Bentley-Gallup recorded the same movement on a different question: the share of 18-to-29-year-olds saying AI does more harm than good went from 27% in 2024 to 36% in 2025 to 47% in 2026, the largest rise of any age group. Gallup's Gen Z survey (1,572 people aged 14 to 29, fielded February 24 to March 4, 2026) found weekly use holding steady, at 48% of Gen Z adults and 56% of K-12 students, while sentiment turned more negative on three of the four emotions measured since 2025. At Marquette University's May 2026 commencement, students petitioned against the choice of an AI advocate as speaker, the petition failed, and graduates booed the speaker.

The youngest cohort points the other way. Luminate's panel has 13-to-17-year-olds net positive on AI across every entertainment type measured, while 18-to-24-year-olds are net negative, and net interest in AI music among Gen Z and Gen Alpha fell from -6% in May 2025 to -16% in November 2025. Globally, Ipsos finds under-35s are the most nervous age group (52%) and the most excited (56%) at the same time.

Guillaume Verdon proposes that decel psyops are a form of warfare by foreign adversaries here: https://x.com/beffjezos/status/2091958298086113548

AI that stopped being called AI

The field has a term for this, the AI effect. Pamela McCorduck documented in her history of AI that each solved problem, from chess to symbolic algebra, was reclassified by critics as something other than intelligence once it worked. Larry Tesler's version is that AI is whatever has not been done yet, and Nick Bostrom's is that a technique stops being labeled AI once it is useful and common enough. Chess engines, optical character recognition, speech recognition and spam filtering are the standard examples. Fraud scoring, camera autofocus, noise cancellation and route planning run on the same kind of models at scale today without an AI label, on the same class of data centers that 71% of Americans oppose having nearby.

Labeling rules are moving in two directions at once. The EU AI Act's transparency obligations, which require AI-generated content to be marked, apply from August 2026. In January 2026, Valve narrowed Steam's AI disclosure requirement to AI the player encounters in the game, after a period in which the count of releases carrying a disclosure went from about 1,000 in all of 2024 to about 8,000 in the first half of 2025, and to 21% of all 2025 releases through October. In Game Oracle's analysis of about 10,000 Steam releases from January to October 2025, games with an AI disclosure averaged 4 user reviews in their first month against 7 for games without one; the author notes that disclosure correlates with other markers of low-budget releases.

One unlabeled system draws the strongest rejection in the Bentley-Gallup task comparison: 69% of Americans say AI does a worse job than a person at recommending whom to hire, above the 68% for driving and 62% for medical advice. The hiring screen scores the applicant without any surface the applicant sees. No published experiment contrasts unlabeled AI that works on the user's behalf with unlabeled AI that acts on the user.

Chatbots for children drew lawsuits, an FTC inquiry and a state law within a year

1. October 22, 2024: Garcia v. Character Technologies, a wrongful-death suit, is filed by the mother of Sewell Setzer III, 14, after months of conversations with a Character.AI companion. Google and Character.AI reach a settlement in principle on January 7, 2026.
2. April to May 2025: Common Sense Media surveys 1,060 teens aged 13 to 17; 72% have used an AI companion and 52% use one regularly. On April 30, 2025, Common Sense and Stanford's Brainstorm Lab rate social AI companions an unacceptable risk for under-18s.
3. August 14, 2025: Reuters reports internal Meta guidelines that permitted romantic roleplay with minors. Meta later adds parental controls that let parents disable teens' AI chats.
4. August 26, 2025: Raine v. OpenAI, a wrongful-death suit, is filed in San Francisco Superior Court by the parents of Adam Raine, 16.
5. September 11, 2025: The FTC opens a 6(b) inquiry into AI companion chatbots at Alphabet, Character Technologies, Instagram, Meta, OpenAI, Snap and xAI, on a 3-0 vote.
6. October 9-13, 2025: Common Sense Media and NORC survey 1,096 US adults; 56% support age verification for AI companions and 96% say children need online protection.
7. October 13, 2025: California's SB 243 is signed after passing the Senate 33-3 and the Assembly 59-1; it takes effect January 1, 2026 and requires companion chatbots to remind minors to take a break every three hours.
8. October 28, 2025: Senators Hawley and Blumenthal introduce the GUARD Act (S.3062), with penalties up to $100,000 per violation; it advances from the Senate Judiciary Committee on April 30, 2026.
9. October 29, 2025: Character.AI announces the end of open-ended chat for under-18s, effective late November 2025; about 10% of its roughly 20 million monthly users are minors.
10. February 12-18, 2026: A National Parents Union poll of 1,511 K-12 public school parents by Echelon Insights finds 80% want more AI guardrails for children.
11. June 8, 2026: Common Sense Media's census of 1,204 children aged 9 to 17 finds 86% use AI, 24% daily, and 44% say no parent has discussed AI safety with them.

Work: rollout meets resistance

Gallup's State of the Global Workplace 2026 reports that in organizations that have deployed AI, 65% of US workers say it has improved their productivity, while 12% strongly agree it has transformed how work gets done. Pew has 71% of US adults expecting fewer jobs over 20 years, and Bentley-Gallup has 79% expecting fewer US jobs over 10 years.

Opt-in worker surveys record active resistance. Writer and Workplace Intelligence (April 2026, 2,400 knowledge workers including 1,200 executives in the US, UK and Europe) found 29% admitting to sabotaging their employer's AI strategy, 44% among Gen Z, with 30% of those citing fear that AI would take their job. Software Finder (June 2026, 1,005 US workers) found half of workers resisting new AI tools, and resisters earning about 20% less than adopters ($65,645 against $81,526). Ivanti (February 2025, 6,000 office workers) found 30% using AI without telling their employer, for fear of job cuts. Glassdoor's May 2026 midyear report counted AI mentions in employee reviews up 240% year on year, with sentiment at 53% negative, from 55% positive a year earlier.

Klarna's AI assistant handled 2.3 million customer conversations in February 2024, the work of about 700 agents by the company's count. In May 2025 the company began hiring human agents again, with CEO Sebastian Siemiatkowski saying cost had weighed too heavily in the decision and service quality had fallen; Klarna later described the human hiring as a pilot and said its AI investment continued.

The industry's account

Wired published "Silicon Valley Doesn't Get Why You Hate AI" on August 20, 2026, and TechCrunch ran "AI was supposed to win people over by now, it hasn't" the day before. Mark Zuckerberg published an essay of about 6,500 words on his AI outlook in August 2026; Dario Amodei responded on X on August 15 that public perception had become a "crisis of trust" and that people suspect the industry is devising new ways to exploit them. Airbnb CEO Brian Chesky attributed the backlash in part to the industry not shipping products that regular people like and in part to how the industry talks about AI. Bentley-Gallup's 2026 data has self-reported knowledge of AI at 70%, up 6 points on 2024, in the same year that harm-over-good rose 8 points.

Where AI is accepted

Ipsos's 2026 AI Monitor (23,532 adults in 32 countries, fielded March 20 to April 3, 2026) has Asia and Latin America net optimistic and Europe and North America net nervous, with excitement and nervousness at near parity worldwide. In 17 of the 28 countries surveyed in consecutive years, fewer people than a year earlier said AI's benefits outweigh its drawbacks. 46% across the 32 countries said they would trust a generative AI tool less if advertisers influenced its answers. In Pew's 25-country study, the share more concerned than excited was highest in the United States and Italy (50%), Australia (49%), Brazil (48%), Greece (47%) and Canada (45%), and lowest in South Korea (16%), India (19%), Israel (21%), Nigeria (24%), Turkey (26%), Japan (28%) and Germany (29%). KPMG and the University of Melbourne (48,340 people in 47 countries, November 2024 to January 2025) found 66% using AI regularly, 83% expecting benefits, 46% willing to trust AI systems and 70% wanting regulation. In China, Ipsos found AI trusted above a human for some kinds of immediate health advice.

Two datasets on AI advertising to young adults give opposite readings. An EMARKETER survey from April 2025 found 57% of millennials and 50% of Gen Z liking an ad or sponsored post more when told it was AI-made. IAB and Sonata Insights surveyed 505 US Gen Z and millennial consumers between October 2025 and January 2026 and found 45% positive about AI-generated ads, against 82% of advertising executives who expected them to be positive, with 39% of Gen Z negative compared with 20% of millennials. Buffer's analysis of 1.2 million social posts found AI-assisted posts at a median 5.87% engagement rate against 4.82% for fully human-written posts.

Sources

1. Pew Research Center, Young adults in the U.S. are increasingly wary of AI (August 18, 2026): https://www.pewresearch.org/short-reads/2026/08/18/young-adults-in-the-us-are-increasingly-wary-of-ai-concerned-it-will-take-jobs/
2. Pew Research Center, Key findings about how Americans view AI (March 12, 2026): https://www.pewresearch.org/short-reads/2026/03/12/key-findings-about-how-americans-view-artificial-intelligence/
3. Courthouse News, Pew: Young Americans grow more wary of AI: https://www.courthousenews.com/articles/pew-young-americans-grow-more-wary-of-ai-as-job-fears-rise
4. Implicator, Young adults' AI concern reaches 55% (Marquette commencement): https://www.implicator.ai/pew-young-adults-ai-concern-55-percent/
5. Gallup, Americans Oppose AI Data Centers in Their Area (May 13, 2026): https://news.gallup.com/poll/709772/americans-oppose-data-centers-area.aspx
6. Gallup / Bentley University, Americans Cool Toward AI (August 2026): https://news.gallup.com/poll/712751/americans-cool-toward.aspx
7. Bentley University, Bentley-Gallup AI research: https://www.bentley.edu/gallup/ai
8. Gallup, Gen Z's AI Adoption Steady, but Skepticism Climbs (April 8, 2026): https://news.gallup.com/poll/708224/gen-adoption-steady-skepticism-climbs.aspx
9. Gallup, State of the Global Workplace 2026, coverage by UNLEASH: https://www.unleash.ai/strategy-and-leadership/analysis/gallups-state-of-the-global-workplace-2026-report-three-key-decisions-for-hr-leaders
10. KPMG / University of Melbourne, Trust, attitudes and use of AI: A global study 2025: https://kpmg.com/xx/en/our-insights/ai-and-technology/trust-attitudes-and-use-of-ai.html
11. Ipsos, Global Attitudes on AI 2026: https://www.ipsos.com/en/global-attitudes-ai-2026-wonder-vs-worry-divide-deepens
12. Luminate, Consumer survey finds increasing discomfort with generative AI in music: https://luminatedata.com/blog/consumer-survey-finds-increasing-discomfort-with-generative-ai-in-music/
13. Music Ally, Luminate study explores Americans' interest in GenAI content (July 27, 2026): https://musically.com/2026/07/27/luminate-study-explores-americans-interest-in-genai-content/
14. NPR, AI music is flooding streaming platforms. But listeners like it less and less (May 2, 2026): https://www.npr.org/2026/05/02/nx-s1-5804489/music-listeners-dislike-ai-music-study
15. ProMarket, Consumers prefer AI music until they're told it's AI (May 4, 2026): https://www.promarket.org/2026/05/04/consumers-prefer-ai-music-until-theyre-told-its-ai/
16. PRS for Music, member AI survey (January 2026): https://www.prsformusic.com/m-magazine/news/prs-for-music-ai-survey-2026
17. PsyPost, People consistently devalue creative writing generated by AI (Raj et al., Wharton): https://www.psypost.org/people-consistently-devalue-creative-writing-generated-by-artificial-intelligence/
18. Nuremberg Institute for Market Decisions, Transparency without trust: https://www.nim.org/en/publications/detail/transparency-without-trust
19. MarTech, Consumers like AI content until they know it's AI (Capgemini, YouGov data): https://martech.org/consumers-like-ai-content-until-they-know-its-ai/
20. Klaviyo, Consumer trust in AI (Gartner 2026 figure): https://www.klaviyo.com/solutions/ai/consumer-trust-in-ai
21. Washington State University, Using the term "artificial intelligence" in product descriptions reduces purchase intentions (Cicek et al.): https://news.wsu.edu/press-release/2024/07/30/using-the-term-artificial-intelligence-in-product-descriptions-reduces-purchase-intentions/
22. British Food Journal, AI-generated image labels and purchase intention: https://www.emerald.com/bfj/article/doi/10.1108/BFJ-08-2025-1101/1332772/
23. PBS / AP, Merriam-Webster's word of the year for 2025 is "slop": https://www.pbs.org/newshour/nation/merriam-websters-word-of-the-year-for-2025-is-ais-slop
24. The Conversation, Slop, vibe coding and glazing: AI dominates 2025's words of the year: https://theconversation.com/slop-vibe-coding-and-glazing-ai-dominates-2025s-words-of-the-year-269688
25. Reuters Institute, Digital News Report 2026, executive summary: https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026/dnr-executive-summary
26. Authors Guild, AI survey on consent and compensation: https://authorsguild.org/news/ag-ai-survey-reveals-authors-overwhelmingly-want-consent-and-compensation-for-use-of-their-works/
27. Windows Central, Game Oracle "AI stigma" Steam study: https://www.windowscentral.com/gaming/pc-gaming/ai-game-development-stigma-study
28. The Conversation, Steam AI disclosure rules: https://theconversation.com/are-video-game-developers-using-ai-players-want-to-know-but-the-rules-are-patchy-274850
29. Data Center Watch, project report: https://www.datacenterwatch.org/report
30. NBC News, Data center opponents have blocked or delayed projects worth nearly $130 billion in 2026 (June 12, 2026): https://www.nbcnews.com/tech/tech-news/data-center-opposition-sharply-rising-2026-study-finds-rcna349728
31. Wikipedia, Opposition to AI data centers: https://en.wikipedia.org/wiki/Opposition_to_AI_data_centers
32. CNBC, AI data center outrage is showing up everywhere from ads to elections (August 20, 2026): https://www.cnbc.com/2026/08/20/ai-data-center-election-backlash.html
33. Fortune, Mark Cuban on the AI backlash (June 26, 2026): https://fortune.com/2026/06/26/mark-cuban-ai-backlash-silicon-valley/
34. Fortune, Nearly a third of workers admit to sabotaging their company's AI (July 30, 2026): https://fortune.com/2026/07/30/nearly-a-third-of-workers-admit-to-sabotaging-their-companys-aiand-smaller-paychecks-may-explain-why/
35. EMARKETER, Employees fear disclosing AI use would threaten their job (Ivanti data): https://www.emarketer.com/content/employees-fear-disclosing-ai-use-would-threaten-their-job
36. EMARKETER, Humans still win the content race (Ipsos and April 2025 ad data): https://www.emarketer.com/content/humans-still-win-content-race
37. CNBC, America's AI backlash: how the effort to keep worker trust is evolving (August 23, 2026): https://www.cnbc.com/2026/08/23/ai-backlash-workers-job-losses.html
38. Forbes, Klarna reverses on AI, says customers like talking to people (May 18, 2025): https://www.forbes.com/sites/quickerbettertech/2025/05/18/business-tech-news-klarna-reverses-on-ai-says-customers-like-talking-to-people/
39. Common Sense Media, Nearly 3 in 4 teens have used AI companions: https://www.commonsensemedia.org/press-releases/nearly-3-in-4-teens-have-used-ai-companions-new-national-survey-finds
40. Common Sense Media, inaugural annual study on AI use by tweens and teens (June 8, 2026): https://www.commonsensemedia.org/press-releases/common-sense-media-releases-inaugural-annual-study-on-ai-use-by-tweens-and-teens
41. FTC, inquiry into AI chatbots acting as companions (September 11, 2025): https://www.ftc.gov/news-events/news/press-releases/2025/09/ftc-launches-inquiry-ai-chatbots-acting-companions
42. US Congress, GUARD Act, S.3062: https://www.congress.gov/bill/119th-congress/senate-bill/3062/text
43. California Legislature, SB 243: https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202520260SB243
44. CNN, Character.AI and Google settle teen lawsuit (January 7, 2026): https://www.cnn.com/2026/01/07/business/character-ai-google-settle-teen-suicide-lawsuit
45. Wikipedia, Raine v. OpenAI: https://en.wikipedia.org/wiki/Raine_v._OpenAI
46. Character.AI, under-18 chat announcement (October 29, 2025): https://blog.character.ai/u18-chat-announcement/
47. TechCrunch, AI was supposed to win people over by now, it hasn't (August 19, 2026): https://techcrunch.com/2026/08/19/ai-was-supposed-to-win-people-over-by-now-it-hasnt/
48. Futurism, Clueless AI CEOs still baffled why everybody's so mad about AI: https://futurism.com/artificial-intelligence/ai-ceos-public-mad-outrage
49. Futurism, The more people learn about AI, the more they want it out of their lives: https://futurism.com/artificial-intelligence/ai-learn-negative-sentiment-polling-american-attitudes
50. Wikipedia, AI effect: https://en.wikipedia.org/wiki/AI_effect
51. Without public links: Wired, Silicon Valley Doesn't Get Why You Hate AI (August 20, 2026); Annenberg Public Policy Center data center poll (2026); Schilke and Reimann, AI disclosure and trust (13 experiments, 2025); IAB / Sonata Insights, The AI Ad Gap Widens (October 2025 to January 2026); Glassdoor midyear report (May 2026); Software Finder worker survey (June 2026); Writer / Workplace Intelligence survey (April 2026); Buffer analysis of 1.2 million posts; National Parents Union / Echelon Insights poll (February 2026); Common Sense Media / NORC poll (October 2025).