Workers Using Banned AI Apps in Majority of Offices

Fact-check confidence: HIGH

All key facts are directly stated in the source with matching numbers and specifics. Survey methodology is clearly described. OpenAI data retention policy is accurately cited from their website. No unsupported claims or vague attributions detected.

Two-thirds of workers use banned AI apps despite employer restrictions

The short version

66% of workers use AI tools their employers have banned
43% have entered emails or work correspondence into public AI platforms
48% of workers faced formal consequences for prohibited AI use
OpenAI retains user data by default to train its models

66%

Workers using banned AI tools at work

43%

Workers who entered work emails into public AI tools

48%

Workers who faced formal consequences for prohibited AI use

1,200+

Office professionals surveyed across US, UK, Australia, and Japan

What happened

PagerDuty surveyed more than 1,200 office professionals at large non-tech corporations and non-profits ($500M+ revenue) and found that 66% use AI tools despite employer bans. Of those workers, 43% admitted entering work emails or correspondence into public platforms such as ChatGPT, Claude, or Gemini. Nearly half faced formal consequences for the prohibited use.

Context

Workplaces have increasingly implemented AI restrictions to protect proprietary information and comply with data security standards. OpenAI retains conversation data by default to train and improve ChatGPT, unless users opt out or use paid enterprise versions—a setting many workers may not know about or consider when entering sensitive materials.

Both sides

Bull

Workers using AI tools claim they understand AI's workplace applications better than managers writing policies, suggesting employee-led adoption could drive beneficial innovation if properly managed.

Bear

Unauthorized AI use exposes companies to data breaches, intellectual property loss, and regulatory violations when workers share proprietary information with third-party platforms outside IT oversight.

In plain words

Lots of workers use ChatGPT and similar AI tools at their jobs even though their bosses said they can't, and some paste work emails into these tools without realizing the company behind them saves that data. This creates a risk because sensitive company information could end up in places the company didn't approve.

What may happen

Company discovers proprietary data in ChatGPT training set

An employee pastes confidential product designs into ChatGPT for quick analysis. Months later, the company realizes similar designs appear in another competitor's work, potentially traced to OpenAI's training data.

scenario - not a prediction

Employer implements stricter monitoring

After learning about widespread unauthorized AI use, a company deploys monitoring software to detect when employees access ChatGPT or similar tools, or blocks them entirely at the network level.

scenario - not a prediction

Employee retraining on approved AI tools

Rather than banning AI, a company implements approved, enterprise-grade AI solutions with better data privacy controls and trains workers to use them safely for specific tasks.

scenario - not a prediction