OpenAI lanza dos modelos de IA de código abierto tras el éxito de la china DeepSeek | Diario Financiero
OpenAI has recently introduced its first open-weight artificial intelligence models in years, marking a strategic shift away from its history of closed software and a response to growing competition and new global regulatory pressures. This decision aims to balance innovation with safety, allowing developers and companies to adapt the systems on their own servers, although the lack of access to training data limits full transparency of the process. This move has significant implications for the open-source movement, as it demonstrates how large corporations attempt to adopt open research and academic standards without relinquishing complete control over their most critical assets. Although the model weights are accessible, the concealment of training data maintains a gap between technical publication and true openness, sparking debate about safety and ethics in large-scale AI deployment. The relevance to open data lies in the fact that this event highlights the current tension between the availability of algorithms and the accessibility of fundamental data. For AI to genuinely contribute to open science and auditability, it is crucial to advance toward models that not only share weights but also ensure transparency in information sources, ensuring that the community can verify and reproduce results in an integral manner.
Source: df.clPublished on 2025-08-07
Related news
- ¡Latinoamérica pierde el control de sus datos! Alerta por falta de soberanía digital - El Imparcial de Oaxaca
- ¿Tus creaciones hechas con IA tienen dueño legal o a quién le pertenecen?
- Reduce Transparencia para el Pueblo desecho de quejas y funciones clave
- OpenAI launches two open weight language models optimised for local use and advanced reasoning - BusinessToday
- OpenAI Returns To Open-source With First Free Models In Six Years