Personal LLM Accounts Drive Shadow AI Data Leak Risks

Netskope’s 2026 Cloud and Threat Report highlights growing risks from Shadow AI as employees increasingly use personal LLM accounts (47% of workplace users), driving a surge in sensitive data exposure and compliance issues; on average, 3% of gen-AI users cause about 223 data-policy violations per month, with heavy adopters seeing roughly 2,100 incidents monthly involving source code, confidential data, and credentials. The report warns attackers could exploit LLM inputs and urges stronger governance, monitoring, and user education to curb accidental leakage, noting some progress as personal-account use reportedly dropped from 78% to 47% year over year.