Expertos analizaron cómo usa ChatGPT los datos públicos de usuarios de Internet, ¿qué encontraron?
A European working group has preliminarily concluded that large language models pose significant risks to fundamental rights when trained on public data that may contain sensitive personal information. This situation demands a delicate balance between the legitimate interests of businesses and the protection of privacy, highlighting the need to implement robust security measures, such as anonymization and the exclusion of specialized categories of data. Transparency does not automatically justify the processing of information, as public availability does not imply explicit user consent for its use in algorithmic training. Therefore, it is crucial to determine whether the individual clearly intended to make their data accessible, reinforcing the principle that technical accessibility does not override legal protections or the inherent rights to individual privacy in the digital environment. This report is relevant to open data because it establishes vital legal precedents regarding the ethics and legality of using public information in artificial intelligence. By defining the boundaries between accessible data and protected personal data, it provides essential guidance to ensure that open data initiatives comply with European regulations and prevent the erosion of individual freedoms in the digital age.
Source: elcolombiano.comPublished on 2024-05-29
Related news
- La relación entre creativos y tecnólogos se tensa con demandas y acuerdos por el uso de datos para entrenar la AI
- NVIDIA responde a demandas por infracción de derechos de autor en entrenamiento de IA
- La vivienda ya está al precio más alto de los últimos 14 años
- The 2 Johnnies challenge 'influencer' label and explain RTE 2FM exit