Más de 16.000 bases de datos quedaron abiertas en internet por un error de configuración: había contraseñas, teléfonos y direcciones
The ease with which artificial intelligence enables the creation of applications exposes a serious security risk: misconfigurations that leave personal data accessible to anyone. Recent investigations have revealed that thousands of databases on popular open-source platforms are publicly exposed, not due to external attacks, but because of human and automated errors in failing to enable basic protections. The problem lies in the fact that AI assistants, which accelerate development through generative programming, often omit the default activation of row-level security. This vulnerability leaves names, addresses, passwords, and other sensitive information visible, affecting everything from commercial services to government institutions, particularly in regions with weak regulatory frameworks. This case is crucial for the open-source community because it demonstrates that technical accessibility does not guarantee security. It underscores the need for developers to rigorously verify privacy configurations and for AI tools to prioritize data protection by default, ensuring that technological innovation does not compromise end-user privacy.
Source: es.gizmodo.comPublished on 2026-10-01
Related news
- Hackers roban millones de datos de personal militar de EE. UU. en una brecha de seguridad prolongada
- Esta es la cronología de los incidentes relacionados con la IA desde el ataque a Hugging Face
- OpenAI enfrenta la primera demanda por acciones autónomas de una IA en un ciberataque
- OpenAI enfrenta una demanda por el ciberataque de sus agentes de IA contra Hugging Face - La Tercera