One in five employees inputs sensitive data into non-secure AI, says report
Accenture’s survey of 1,000 Irish office workers found 19% input sensitive business data (customer/financial) into free, unsecured AI tools. Despite frequent training, confusion persists around reporting suspicious messages, and deepfake/AI‑phishing risks are top concerns—highlighting cultural and process gaps.
CORTEX Protocol Intelligence Assessment
Business Impact: Shadow‑AI usage creates data leakage and compliance risks. Technical Context: Insufficient guardrails and unclear responsibility dilute reporting and response.
Strategic Intelligence Guidance
- Deploy sanctioned enterprise AI with DLP and logging.
- Enforce policy controls blocking unsanctioned AI sites.
- Run role‑based training on AI misuse scenarios.
- Establish a clear incident reporting path for AI‑related events.
Vendors
Threats
Targets
Intelligence Source: One in five employees inputs sensitive data into non-secure AI, says report | Oct 22, 2025