Pressure pushes US workers to skip AI & email checks

Trustmi has published survey findings suggesting many US workers are bypassing checks and using unapproved AI tools under workplace pressure. The research highlights growing concern about how urgency affects employee judgment.

The survey of 1,000 US adults, including 591 employed respondents, found that 53% had used an AI tool without being sure it was approved because they felt pressure to work faster or do more with AI. It also found that 71% had noticed an error or suspicious detail in an email, document or work request only after they had already responded, shared it or acted on it.

Among employed respondents, 43% said they had discovered those warning signs after the fact on multiple occasions.

Pressure and checks

The findings suggest time pressure is weakening basic verification habits in day-to-day work. Some 84% of employed respondents said a heavy workload makes it harder to pay attention to important or sensitive tasks, while 83% said tight deadlines have the same effect.

A further 68% said they were likely to spend less time checking whether a request was legitimate when they felt pressure at work. That suggests the issue goes beyond fraud awareness and affects how workers handle routine approvals and communications.

Trustmi, which focuses on payment security, said the results show the limits of relying on employees to spot suspicious details in fast-moving workflows. It argues that pressure linked to productivity goals and AI adoption is creating conditions in which mistakes and risky behaviour are more likely.

Shai Gabay, Co-founder and Chief Executive Officer of Trustmi, said the problem affects a wide range of business processes. "From everyday emails to payment invoices, many organizational checks on accuracy are human driven. But individual and industry pressures can drive an urgency that changes the way humans make those judgment calls, and this creates an opening for risk," Gabay said.

The survey also points to a gap between seeing warning signs and spotting them in time to prevent harm. In practice, that can mean staff approve, reply to or circulate requests before noticing inconsistencies.

AI and shortcuts

The data also suggests pressure to show gains from AI is affecting software choices inside organizations. More than half of employed respondents said they had turned to an AI tool without being sure it was approved.

Among that group, 13% said they had done so frequently, 27% occasionally and 13% once. That suggests the use of unsanctioned tools is not limited to one-off experimentation.

For companies, the issue is likely to raise governance questions as well as security concerns. Unapproved tools can create uncertainty over how information is handled, who has access to it and whether standard controls are being followed.

The results come as businesses continue to embed AI into routine work while also demanding faster output from staff. In that environment, verification steps that depend on careful reading or manual review may be easier to skip.

Trustmi's work centres on payment fraud and payment errors, where realistic emails, altered invoices and socially engineered requests can lead to financial loss if staff act too quickly. It said overworked employees are especially exposed when requests appear plausible and arrive through trusted channels.

Gabay said speed demands can directly weaken safeguards. "When employees are asked to work faster, this can come at the expense of verification controls that prevent errors and fraud alike. For critical areas like payment workflows, it's simply not enough to rely on workforces constrained by demanding priorities to catch every mistake - extra protection is essential," he said.

Call for safeguards

Respondents also signalled support for additional checks beyond human review. According to the survey, 87% of employed respondents said it would be valuable to have technology automatically perform a second check when a workplace request contains unusual details or other warning signs.

That reflects a broader shift in how businesses are approaching operational risk. As AI-generated content becomes harder to distinguish from legitimate communications, companies may face growing pressure to add automated controls around approvals, document handling and payment processes.

The survey was conducted by Dynata and examined how urgency and competing demands shape attention, verification and responses to potentially suspicious requests. It suggests the challenge is not simply whether workers can recognise a red flag, but whether they can do so in time to stop an action already under way.

For finance and security teams, that timing issue may prove central as companies weigh the trade-off between speed and control. The survey's clearest numerical finding was that 71% of employed respondents had identified an error or suspicious detail only after they had already responded to, shared or acted on an email, document or work request.