Americans Report High Skepticism Toward AI
WordPress VIP's 2026 "Future of the Web" survey of 1,200 U.S. adults finds 86% of consumers always or sometimes check the original source after receiving an AI-generated summary. The vendor-commissioned research further reports that 42% of respondents rank AI answers without clear attribution as the least trustworthy content they encounter online - above confusing medical bills (15%), airline fees (19%), and opaque privacy policies (24%). Separately, 60% say AI in brand messaging is a turnoff, and 80% say online information should remain openly accessible rather than controlled by a small number of organizations. The research covered 2,000 respondents including 800 enterprise decision-makers, was commissioned by WordPress VIP (an Automattic company), and was published June 16, 2026. The survey also reports 75% of respondents prefer human customer service representatives over AI and 56% say they can distinguish AI from human interactions.
What happened
WordPress VIP's 2026 "Future of the Web" report, based on a survey of 2,000 respondents including 1,200 U.S. adults and 800 enterprise decision-makers, finds that 86% of consumers always or sometimes check the original source after receiving an AI-generated summary. The vendor-commissioned research, published June 16, 2026, also reports that 42% of consumers rank AI answers without clear attribution as the least trustworthy content they encounter online - above confusing medical bills (15%), airline fees (19%), and opaque privacy policies (24%). Per the same report, 60% of consumers say AI in brand messaging is a turnoff, and 80% say online information should remain openly accessible rather than controlled by a small number of organizations.
Customer service findings
The survey reports 75% of respondents find human customer service representatives more helpful than AI, while 15% prefer AI interactions, and 56% say they can confidently distinguish AI from human chats.
Editorial analysis
The source-checking behavior documented in the survey - 86% going back to verify AI-generated summaries - is directly relevant to practitioners building AI-facing content products. Systems that surface AI answers without clear provenance impose an additional verification step on users and increase abandonment risk. The 42% distrust figure, ranking AI without attribution as less trustworthy than medical bills, airline fees, or privacy policies, signals a concrete friction point that content attribution and structured-data approaches can address. Practitioners should note the vendor-commissioned nature of the research: WordPress VIP is an enterprise CMS provider that benefits commercially from heightened AI-attribution concerns, and findings should be weighed accordingly.
What to watch
- •Uptake of provenance and citation features in content generation platforms
- •Transparency labels or watermarks adopted by major publishers and platforms
- •Policy or industry standards addressing concentrated control of information access
- •Vendor claims about verifiability and audit trails for generated content
Key Points
- 1WordPress VIP's 2026 survey finds 86% of consumers check the original source after AI summaries, raising the bar for provenance in AI-facing content products.
- 242% of respondents rank AI answers without attribution as the least trustworthy content online - more distrusted than medical bills (15%), airline fees (19%), or privacy policies (24%).
- 3Findings are vendor-commissioned (WordPress VIP, an enterprise CMS company) and should be read alongside that commercial context; the trust gap still signals real design friction.
Scoring Rationale
A vendor-commissioned survey (by WordPress VIP, an enterprise CMS provider) finding that 86% of consumers cross-check AI summaries and 42% rank unattributed AI answers as their least-trusted content type is directly relevant to practitioners designing AI-facing products and attribution workflows. The vendor origin and specialist-blog distribution modestly limit the score relative to independent research on AI trust.
Sources
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems