‘Over-trust’, ‘under-scrutiny’ leading to AI mistakes – and more work for managers
Seven in 10 U.S. managers say employees have made mistakes using artificial intelligence tools in the past year, with some errors costing employers more than US$50,000, according to a new survey by Resume.org.
The January 2026 poll of 1,146 managers found that 70% had observed at least one AI‑related error from a direct report over the previous 12 months. These incidents are often repeated: 12% of managers reported seeing such mistakes “many times”, while 43% said they had seen them “several times”.
Resume.org describes this pattern as an “AI slop” crisis, referring to low‑quality or unchecked AI output that finds its way into completed work and triggers business consequences. For HR professionals, the findings highlight that AI risk now sits squarely in day‑to‑day people management, not just in IT.
“Most AI-related mistakes stem from over-trust and under-scrutiny,” says Kara Dennison, head of career advising at Resume.org. “Employees treat AI outputs as finished work rather than as a starting point. AI is reliable when used as an assistant, not a decision-maker. Without human judgement and clear processes, speed becomes a risk, and efficiency gains can turn into costly mistakes.”
Who's making mistakes with AI?
Managers were also asked which age groups they saw as most likely to make AI‑related mistakes. Thirty‑four per cent identified Gen Z employees (aged 18 to 29) as the most error‑prone group, followed by Millennials (30 to 46) at 26%. Gen X (47 to 61) was cited by 18% of respondents, and Baby Boomers (62+) by nine per cent. Twelve per cent reported no clear difference by age.
“Younger workers aren’t necessarily more careless, but they’re often using AI more frequently and earlier in their workflows,” Dennison said. “There is also a training gap. Organisations often assume younger employees intuitively understand AI, yet provide little guidance on verification, risk, or appropriate use cases. As a result, AI may be treated as an answer engine rather than a support tool.”
Despite the widespread adoption of AI and positive developments that have come with it, one in five of the white‑collar professional workers globally say they have encountered misinformation, errors, or misleading outputs from AI tools, according to a previous report.
Inaccurate information, missing context
The most common problems flagged by managers involve incorrect facts and missing context, according to the Resume.org report. Among those who had seen AI‑related mistakes, 58% reported staff submitting work that contained factual inaccuracies generated by AI tools. More than half said they had seen errors where AI failed to reflect important contextual factors, nuance or constraints.
Other frequently cited issues included low‑quality content (41%) and poor recommendations (35%). More than one‑third of respondents reported communication problems, such as unclear or inappropriate messaging, tied to AI usage.
The survey also found that 29% of managers had encountered AI‑related mistakes that raised confidentiality, privacy or compliance concerns, and 18% had seen AI worsen conflicts or sensitive situations. For HR, these patterns point directly to the need for role‑specific AI guidance and training, especially in high‑risk functions such as HR, legal and client‑facing roles.
According to Resume.org, the findings underscore the limits of current AI tools in handling “complex, nuanced, or context-dependent tasks”, even as they accelerate drafting and information retrieval.
For many organisations using AI, every 10 hours of apparent efficiency may be dampened by nearly four hours that disappear into fixing mistakes and filling gaps, according to a previous report.
Workflow disruptions and relationship damage
AI mistakes are not contained to the employees who make them, the research suggests, according to Resume.org. A majority of managers (58%) said they had personally been affected by AI‑related errors made by direct reports.
Within organisations, 44% of managers reported that co-workers had been affected by such errors, while 24% said superiors had experienced negative consequences. Outside the organisation, nearly 40% of managers reported that clients had been negatively affected by AI‑related mistakes, and 20% cited impacts on vendors and suppliers.
“Many organizations adopted AI faster than they set clear guidelines or training, leaving employees unsure when to rely on it and when to challenge it,” Dennison said. “AI works best as a support tool. Problems arise when it replaces human judgment instead of reinforcing it. Employees, especially early in their careers, may not yet have the expertise to spot subtle inaccuracies or flawed recommendations.”
For HR leaders, the breadth of those impacts signals that AI governance now intersects with culture, trust and employer brand.
Extra work, missed deadlines and direct financial losses
The survey indicates AI‑related mistakes are driving both workflow inefficiencies and broader business risks.
Fifty‑nine per cent of managers said they personally had to spend extra time correcting or redoing tasks affected by AI errors. A further 53% reported that their direct reports had to redo work, and 45% said other co-workers were drawn into remediation efforts.
Missed deadlines were reported by 25% of managers. In addition, 28% cited damage to credibility or brand, while 18% pointed to lost opportunities linked to AI‑related mistakes.
“To get the most benefit from AI chatbots, employees should use them as accelerators, not decision-makers. AI works best for drafting, summarising, and exploring options, while humans remain responsible for validation, context, and final judgement,” Dennison said.
Beyond indirect costs, the survey documents direct financial impacts. Nearly one in five managers said AI‑related mistakes had cost their business more than US$10,000, and 5% reported losses exceeding US$50,000. Resume.org says these figures “provide a compelling business case for investing in AI risk management, including preventive measures, training, and oversight”.
“Unchecked, inaccurate or hallucinated AI content can damage trust and credibility among colleagues and clients and lead to reputational damage. There could be IP risks for the employer too if the AI-generated content is not original and derived from other sources,” say Hannah Mahon, partner in the employment, labour and pensions group, and Rebecca Denvers, principal associate professional support lawyer, both at Eversheds Sutherland.
“A crucial part of addressing this comes from training employees on proper AI use. But organisations must have measures in place to protect employees and themselves if things go wrong.”
Employers that cut staff as they rolled out AI may soon be hiring many of those workers back, according to recent predictions that suggest AI-triggered layoffs have overshot.
