In this seminar, Ferran Marqués, professor at the Polytechnic University of Catalonia with 30 years of experience in image processing and computer vision, addressed the latest developments in Visual Intelligence (AI). The talk focused on the technological and social challenges that these technologies present in terms of traceability and reliability.
Computer vision, an area of Artificial Intelligence, is responsible for processing, analysing and interpreting digital images using Deep Learning technologies. Advances in this area have had a significant impact on fields such as automotive, medicine and satellite imagery. The exponential development of artificial intelligence in recent years has been due to three factors: the explosion of data, the development of powerful algorithms to process this data, and the availability of computers capable of supporting these algorithms and the large amount of data they process.
In this context, programmes known as "diffusion models" have emerged that have the ability to automatically generate images from textual descriptions. However, this ability to create images from databases poses the risk of losing photographic authenticity, which opens up the possibility of disseminating false images with malicious intent. The trustworthiness of an image becomes a questionable issue, which calls for the development of new technologies to verify its authenticity in photographic terms.
In addition, these technologies also raise concerns about human, social and even ecological costs. In terms of both human costs, especially in relation to possible labour exploitation in countries where labour is cheaper. The training of artificial intelligences demands huge databases of images categorised by text, a process that can only be manually refined by humans. This situation raises ethical and social dilemmas that need to be addressed in the development and application of these technologies.
Finally, the question arises as to whether we can be confident that these technologies will not pose a risk in the future. Ferrán Marqués gives us his view on the latest technological advances used to verify the origin of these images. Importantly, however, Ferrán also highlights the risk that the development of artificial intelligence carries in terms of partially or totally replacing certain professions, including medicine. This transformation poses important challenges and reflections on the impact of AI on the labour market and society in general.
Back to the list
