Inteligencia artificial genera imágenes sin afectar derechos de autor
Adobe has introduced Firefly, its generative artificial intelligence model focused on creating images and designs from text, differentiating itself from competitors like DALL-E or Midjourney through its strict respect for intellectual property rights. Unlike other platforms that have faced criticism for using unlicensed data, Firefly is trained exclusively on royalty-free content, specifically licensed material, or images from Adobe Stock, ensuring that it does not infringe upon trademarks or the creativity of other authors. This ethical policy represents a crucial advancement for the sustainability of the digital creative ecosystem, as Adobe plans to implement a system of financial compensation for artists whose works contribute to the model's training. By establishing a framework of fair collaboration, the company not only mitigates legal risks associated with plagiarism but also promotes a standard of transparency and recognition toward original creators, distinguishing its commercial offering from tools that operate in a legal gray area. The relevance of this article to the open data movement lies in its direct challenge to opacity in data collection. While many AI systems use opaque or pirated datasets, Adobe proposes a model where the quality and legality of training data are prioritized, incentivizing the use of licensed and transparent datasets. This approach suggests a future where data used to power artificial intelligence must come from legitimate and consensual sources, fostering responsible practices that balance technological innovation with the moral and economic rights of creators.
Source: infobae.comPublished on 2023-04-09