#4 - All the reasons why we should fear AI - Savoirs en Société - World Latest Movie, Series, Videos and Trailers

#4 - All the reasons why we should fear AI

802 views

Published on Oct 26, 2024

About :

In this video, the fourth episode of an 8-part course for students of the various masters at HES-SO (Switzerland), Richard-Emmanuel Eastes examines the complex issues raised by AI and breaks them down into different ethical and societal challenges, from the most global to the most specific. He begins by recalling certain fundamental concepts covered in previous courses, such as algorithmic biases and flaws in AI design, which can result in stereotypical productions or decisions influenced by incomplete data. The focus is on training biases, the anthropomorphization of machines, and the effects of limited diversity in digital professions. By exposing the risks linked to the illusion of humanity in chatbots, particularly using the example of online pornography, he highlights the dangers of overconfidence in systems designed to appear human. The course also explores the non-determinacy of AIs, highlighting the challenges posed by neural networks and their partially opaque nature, thus raising crucial questions about the interpretability and transparency of decisions. The scalability of AI is also addressed, with a debate on weak AI (specific to certain tasks) and the hypothetical strong AI, endowed with general intelligence. Richard-Emmanuel highlights the problem of the alignment of AI, emphasizing the need to program these systems in such a way that they pursue objectives in accordance with human values, an issue all the more critical in the perspective of the existence of super AIs. The metaphysical and ethical implications of AI are also discussed, notably in terms of social justice and the impact of AI on human activities such as art and science. The course continues with an analysis of the techno-scientific and macroeconomic issues of AI, examining in particular the industrialization of AI and its dependence on semiconductors, as well as the efforts of large companies to meet the growing energy needs of data centers, with examples such as Microsoft's investments in nuclear energy. The environmental impacts of AI, direct and indirect, are discussed, from the energy effects of training AI models to the increase in extraction and production processes in industry. Finally, the course highlights the social and political issues, addressing the exploitation of workers in digital labor and the risks to democracy, the manipulation of public opinion and the weakening of truth criteria. On the pedagogical level, Richard-Emmanuel concludes on the importance of guaranteeing equitable access to AI tools, training students in their responsible use and developing their analytical skills. This course thus offers essential keys to understanding the profound and often contradictory issues raised by AI, and to considering its impacts on our lives and societies. Test your knowledge on the video with Wooflash: https://app.wooflash.com/join/HNAP6DA...

vidoe Comments

Trend Videos