How Low-Code AI Tools Created a Data Exposure Time Bomb

A major security researcher has uncovered a critical gap in AI-powered app development: thousands of applications built with platforms like Lovable and Replit lack any meaningful security or authentication, leaving corporate secrets and personal data exposed to anyone who finds the URL. With about 40% of these apps containing sensitive information—from medical records to financial data to customer conversations—this discovery exposes a fundamental tension between speed of development and security by design. The incident forces a reckoning with how AI development tools are lowering barriers to entry without enforcing security baselines.