• by Laurie Sullivan , Yesterday

Consumers trust independent third-party sources much more than they trust AI or brands when a brand says one thing and an AI chatbot says another. When facing this situation, consumers do not automatically believe either source, and they end up doing their own research to find the truth.

Only 29% of consumers trust the brand when its information conflicts with AI-generated information, while just 12% trust the AI answer and 54% look for external validation instead, according to new survey data from Skyword, an enterprise content marketing agency.

“AI tools depend on the information that already exists online to return answers to queries,” Andrew Wheeler, CEO at Skyword, told MediaPost. “Even with safeguards, they can still surface content that is conflicting, incomplete or outdated. That’s where you’ll see trust start to breakdown.”

Wheeler said the data shows consumers check multiple sources instead of taking either the brand or AI answer at face value. For brands, the work starts with understanding where buyers are likely to look for confirmation and to ensure the information they find is clear, accurate and aligns across all types of media.

Skyword’s survey was conducted in April by Dynata, a first-party data platform, and included 1,000 online responses from U.S. adults age 18 and up.

Marketers must have a clearer understanding of how their brands appear in AI-generated answers and citations, as well as how their brands and products are described and positioned in conversations. 

“One inaccurate reference can create real problems, especially if there is not enough stronger information around it,” Wheeler said. “Brands should be looking at the full content ecosystem AI may draw from, including owned content, earned media and third-party sources.”

And from there, it’s crucial to improve the accuracy of their own content, make expert perspectives easier to find, and build more consistent validation across credible outlets, Wheeler said, with a goal to make it difficult for one incorrect mention to carry too much weight about the brand or its products and services.

Brands cannot control all the responses that AI engines throw out because of queries, but the survey found that AI-generated brand information is already driving meaningful consumer behavior.

There is a risk in publishing too much AI-generated content, the study found, with 30% of respondents saying they would be less likely to engage with or buy from a company if they suspected its content was AI-generated, and 86% who said companies should be required to disclose when content is AI-generated.

The study also found that 47% have avoided a purchase, switched brands, warned others, or contacted the company based on what an AI tool told them.

In addition, 17% switched brands because of AI-generated information, 19% avoided a purchase based on what AI told them about a brand, and 46% of full-time employed Americans have made a major purchase or important decision based primarily on AI recommendations.