Read on Twitter
THREAD on AI, ChatGPT, and the Scientific Method. 1/ We've undergone the Scientific Revolution in a (problematic) attempt to understand the world and nature, including human nature, and then to dominate it through science and technology.
2/ The ultimate (again, problematic) goal of this endeavour was to gain control over these elements, to dominate them and bend them to our will. One of the fundamental aspects of the scientific method is the experiment, as it allows for replicability. #ScientificMethod
3/ Once you identify a principle, you can repeat the experiment with the same results and then apply those results on a large scale in the real world. What puzzles me about the current movement towards AI is that we want to create increasingly independent systems
4/ Almost by definition, a crucial goal of developers is not being able to understand these systems. The chief aim is to have systems that function on their own without human input and, therefore, without human control. #ArtificialIntelligence
5/ If you listen to the programmers of AI systems, they often tell you that they do have not a full idea of what happens within these systems. So we have moved from a fundamentally sceptical idea of science and technology founded on replicability to a more faith-oriented approach
6/ We let these systems work, trusting that they will evolve, but without having control over them. This is different from saying, for example, we don't fully understand quantum mechanics, but we know it works and therefore use it without waiting to fully comprehend it.
7/ Here, we don't truly know if these systems work and, once evolved, how they will function. This shift from scepticism and replicability towards a faith-based approach in AI raises vital questions about the future of science, technology, and our role in their development END
• • •
Missing some Tweet in this thread? You can try to
force a refresh
