Shadow AI at Work: 40% of Australian Employees Are Sharing Customer Data with Public AI Tools

A new survey from PagerDuty reveals a widespread and largely hidden pattern of unauthorized AI use inside Australian businesses — with data security implications that are hard to ignore.

The survey, conducted by Wakefield Research across 250 Australian office professionals at companies with at least $500 million in annual revenue, found that 40% have entered customer data into public AI tools such as ChatGPT, Claude, or Gemini. Another 28% have input financial information or disclosed confidential company documents and strategies.

The Scale of Shadow AI

The findings paint a picture of AI adoption that runs well ahead of corporate policy:

• 70% of Australian respondents used AI tools at work despite believing it was not allowed under company policy
• 53% faced formal consequences — warnings or disciplinary action — as a result
• 36% said they would hide their AI use to avoid scrutiny from managers or leadership
• 37% have entered work emails or correspondence into public AI tools outside company systems

Senior leaders were actually more likely than mid-level staff to believe they understood AI better than their company's own AI team (77% vs. 66%).

Policy Gaps and Perception Problems

Part of what is fuelling unauthorized use is a perception that AI policies are applied unevenly. While 82% of respondents believe AI policies exist at their company, 83% also believe there are different rules for leadership compared to everyone else.

Adoption is largely driven by personal use spilling into work: 89% of workers who use AI professionally first tried it in their personal lives, and once they bring it to work, 76% of their total AI time shifts to work contexts.

The retention risk is real too — 72% said they would likely look for a new job that offered better AI skills development.

Read the full article on SMBtech