AI use in email marketing is spreading rapidly, but automating even messages where customer relationships matter can undermine trust, a warning said.

TechRadar reported on Sept. 30 that in Litmus' "State of Email 2026" survey, 28 percent of marketers said they had deeply integrated AI into email tasks and 34 percent said they used it regularly. Only 5 percent said they did not use AI at all. Some 78 percent of respondents said it takes 3 days or less from creating a campaign to sending it, suggesting automation has helped speed up work.

AI is used to draft subject lines and body text, generate images, optimize sending times, analyse the risk of customer churn and design automated sequences. But the broader the automation, the greater the risk of sending incorrect discount rates, end dates or target customers at scale. Google recommends that bulk Gmail senders keep spam rates, based on user reports, below 0.1 percent and not reach 0.3 percent or more.

Consumer trust is also a variable. In the survey, 40 percent of U.S. consumers said trust could fall if they learned a retailer email was written by AI. Another study of 1,100 U.S. office workers found more than 80 percent accepted a message as authentic when a boss used AI for light editing, but the share fell to 40 to 52 percent as the level of AI involvement increased.

There are legal risks, too. Washington state revised its law this year to allow liability if a commercial email subject line contains false or misleading content and the sender knew it, or could have known it based on objective circumstances. The fact that AI wrote it does not remove the sender's responsibility to verify.

A practical benchmark suggested is to use AI for drafting, data analysis and automating repetitive tasks, while having humans recheck key information such as prices, discount rates and promotion end dates. An analysis said limiting AI use to light editing is better for maintaining trust, especially for thank-you or relationship-management emails sent to key customers.