ChatGPT y la privacidad en línea: cómo desactivar el historial y proteger los datos personales
The recent addition of an option to disable chat history in ChatGPT represents a significant step forward in users’ autonomy over their personal data. This setting ensures that interactions are not used to improve the model, guaranteeing their deletion after a short retention period for abuse monitoring. This change responds to the growing demand for transparency and privacy protection, setting a key precedent for the ethical handling of user-generated information on artificial intelligence platforms. However, the relevance of this news extends beyond individual privacy, touching the core of the quality and ethics of the open data used to train these systems. Recent research has revealed that AI models are fed by massive internet datasets that include biased content, copyrighted material, and, in some cases, non-public confidential information. This raises questions about the integrity of the data that underpin these tools’ knowledge, exposing legal and ethical risks stemming from the use of non-consensual or private information. This article is essential for the open data community because it illustrates the inherent tension between the massive availability of information and its responsible use. It highlights the urgent need for clear standards defining which data should be considered public, safe, and appropriate for automated training. By linking technology with legal and copyright implications, the text underscores that open access should not become a channel for the unregulated exploitation of information, calling for greater rigor in the governance of the datasets that support modern artificial intelligence.
Source: semana.comPublished on 2023-04-27