Corporate AI misuse fuels knowledge decay and security risks

A Harvard Business Review study warns that excessive reliance on generative AI can cause a "knowledge decay" in firms. When AI‑generated content is accepted without verification, errors multiply, employee skills erode and trust in internal processes falls. Companies are already creating dedicated roles to audit AI output and limit its use to tasks where it adds clear value.

ESET’s analysis of Hungarian enterprises highlights a related danger called "shadow AI," where employees use unapproved AI tools such as ChatGPT, Gemini or Claude. In 2023 Samsung engineers uploaded confidential code to ChatGPT, causing data leakage that could not be retrieved, and Amazon took similar steps after detecting AI‑generated content resembling its own proprietary data. The report stresses that unchecked AI use creates data‑privacy breaches, GDPR violations, and flawed business decisions, urging firms to establish clear policies, monitoring, and employee training.

Both pieces underline that unchecked AI integration can compromise productivity, security and regulatory compliance, prompting a need for governance frameworks in corporate environments.