UK consumers trust AI marketing less, Optimizely finds

Optimizely has published research showing that 62% of UK consumers trust marketing content less since brands began using AI. The survey highlights a gap between consumer sentiment and how marketers are using the technology.

The study surveyed 1,000 UK consumers and 100 UK marketers. It found that 66% of consumers believe brands are prioritising quantity over quality, while 59% said marketing content quality has declined in recent years.

The findings add to concerns among brands and agencies that the rapid use of generative AI in advertising and marketing may be weakening consumer confidence. More than half of consumers surveyed, 57%, said they can usually tell when content has been generated by AI, while 67% said AI-generated images of people in marketing materials and adverts are creepy.

Consumers also signalled a preference for content that feels less machine-made. The research found that 72% are more likely to engage with content that feels human rather than automated.

Marketers painted a different picture. Among those surveyed, 84% said AI helps reduce manual work during major campaigns, suggesting brands are adopting the technology mainly to improve workflow and output rather than reshape messaging.

That divide reflects a wider tension in marketing teams. AI tools can produce copy, imagery and campaign variations at speed, but the survey suggests audiences may be becoming more sensitive to content they see as generic or formulaic.

Trust gap

The results indicate that trust is becoming a central issue in AI-led marketing. If consumers believe volume is replacing quality, brands may find it harder to persuade audiences that automated content still reflects care, judgement and relevance.

At the same time, marketers appear to be refining how they use the technology. The survey found that 43% see better integration of AI with marketing tools as the biggest opportunity to improve efficiency and optimisation.

Training is also becoming a focus. Over the past 12 months, 42% of marketers said they had invested in training to improve AI outputs, and 44% said they plan to do so over the next 12 months.

This suggests many teams now see AI less as a one-off software purchase and more as an operational shift that requires oversight, process and judgement. Better results may depend not only on the tools themselves, but also on how clearly teams understand where automation helps and where it risks undermining authenticity.

Tara Corey, Senior Vice President of Marketing at Optimizely, said the findings should prompt marketers to reconsider how they deploy the technology.

"AI has made it incredibly easy to produce more content quickly, and I think as marketers we've all been a little guilty of getting caught up in that. When you're under pressure to keep up, it's tempting to see more as 'good', but these findings are a reminder that customers don't necessarily see it that way. Instead of using AI to simply churn out content, we need to be more thoughtful about what we're using it for. AI can take manual work off marketers' plates and give them more time to understand their customers and do their best thinking. But marketers still need to make the call on where AI adds value and where the human touch matters. Consumers aren't looking for more marketing, they're after marketing that's worth their time," Corey said.

The research comes as companies across sectors expand their use of AI in customer-facing work, including website copy, product descriptions, email campaigns and digital advertising. While the technology offers clear gains in speed and lower production effort, the survey suggests consumer reaction may depend less on whether AI is used at all and more on whether the end result feels impersonal.

For marketing departments, that creates a practical challenge. Producing more content can help fill channels and support campaign schedules, but if audiences view that material as lower quality or less trustworthy, the commercial benefit becomes less certain.

The results suggest that, for many consumers, the line between useful automation and unwelcome artificiality is already becoming easier to spot.