Trust in AI content falls to 34% as brands hide the bot

By James Dore · 2026-09-17

What the numbers say and which studies are behind them

Three separate research efforts landed between April and August 2026 and they all found the same reflex. Consumers tolerate AI in principle. They oppose being misled about it.

The sharpest single figure comes from AnswerConnect's annual consumer survey, published in April 2026. 57 per cent of consumers said their trust in a business would decrease if that business relied predominantly on AI for customer service, up from 53 per cent the year before.^[1] That four-point shift matters because it moves in one direction as AI adoption accelerates.

70 per cent of respondents said customer service would get worse if humans were replaced by AI entirely, up from 65 per cent the prior year, and 73 per cent said they would be more loyal to companies that use real people for all service interactions, up from 69 per cent.^[1] That last number is the commercial consequence: loyalty tracking against staffing decisions.

AnswerConnect CEO Natalie Ruiz said that businesses are racing to automate, but customers are pushing back, and people aren't just frustrated by AI, they're actively rejecting it in favor of human interaction.^[1]

The disclosure problem: why impersonation is the specific trigger

ServiceForge's May 2026 survey sharpened the diagnosis. When respondents were asked whether AI should pose as a human in customer service interactions, 85 per cent said no.^[2] That is a rejection of deception specifically, not a rejection of chatbots or automated workflows.

One caveat: ServiceForge did not disclose its sample size or detailed methodology in the available source material. The 85 per cent figure sits directionally consistent with the AnswerConnect data, but readers should weight it accordingly.

A chatbot that identifies itself as a chatbot sits in a different category, commercially and ethically, than one that adopts a human name and persona without disclosure. The consumer data does not penalise the former. It penalises the latter.

Ruiz said that in the race to automate customer experience, many may be eroding the very thing that fuels long-term growth, which is customer loyalty driven by personal connection.^[5] Satisfied customers who feel deceived do not come back.

Gen Z and advertising: how the Bentley-Gallup finding connects

The same consumer reflex showed up in advertising. The Bentley-Gallup study, published 19 August 2026, measured attitudes toward AI in brand communications and found the numbers were worse among the youngest adult cohort.

66 per cent of 18-to-29-year-olds said they hold negative views of businesses that use AI in advertising, and 73 per cent said it was unacceptable for businesses to use AI to create people or voices in ads.^[4] Synthetic spokespeople and AI-generated voices are the specific flashpoint, not AI-assisted editing or AI-generated copy that a human presents.

Across all three studies, the mode of deployment that triggers consumer resistance is the one that obscures the boundary between human and machine. Show the seam and the reaction softens. Hide it and loyalty numbers move.

A fourth study, from PNAS Nexus published in March 2026, adds a related but distinct data point. Only 34 per cent of AI-generated fake news headlines were rated as mostly or completely real by audiences.^[3] That figure measures credibility of AI-generated misinformation, not brand communications or service interactions, and should not be read as a direct trust measure for branded AI content. It does reflect how calibrated public scepticism toward AI-generated text already is, a baseline brands are working against every time they push undisclosed automation into customer-facing channels.

Automation saves money. Undisclosed automation costs loyalty. The three studies together put numbers to that trade-off: a four-point annual drift in trust, an 85 per cent majority against impersonation, and nearly three in four young adults rejecting synthetic people in ads. The Bentley-Gallup study was published 19 August 2026 and the AnswerConnect survey was conducted in April 2026.

James Dore

James Dore is Strategy Director at Bushnote, a leader in SEO and AEO in Australia. He writes on search, AI and the technology decisions shaping government and industry.