Anthropic (which makes the Claude AI system) has released a substantive report (55 page PDF) and some supporting blog posts (one, two) reporting on trends in AI usage across disciplines and in different places. Though the focus is the new set of 'economic primitives' suggested in the report (chapter 2) there's a lot for educators to consider. The new primitives are: task complexity, human and AI skills, use cases (ie., work, learning, home), AI autonomy, and task success. The data shows relations among them. For example, AI is used more for tasks where the expected level of education is higher, but it's rate of success on these tasks declines. The use of Claude for augmentation is more important than the use of Claude to replace humans altogether. There is a "very high correlation between human and AI education... this highlights the importance of skills and suggests that how humans prompt the AI determines how effective it can be." All of that said - the data in this report is complex and the relationships are often subtle, and I would resist articles that try to boil it all down to one or two simple relationships.
Today: Total: [Direct link] [Share]
