Difference between revisions
ai-alignment
Status and infobox
- status:
—→ open - kind:
—→ problem - year:
—→ 1960 - statement:
—→ How do we make capable AI systems reliably pursue the goals their users and society intend?
Source +15 −0
+---
+title: AI alignment
+kind: problem
+status: open
+year: 1960
+statement: How do we make capable AI systems reliably pursue the goals their users and society intend?
+space: machine-minds
+---
+Current methods, human feedback and constitutional training, work in practice and come with no guarantees. The problem grows harder as systems become more capable than the people overseeing them.
+
+## Sources
+
+- [Wikipedia: AI alignment](https://en.wikipedia.org/wiki/AI_alignment)
+
+[[Space:machine-minds|Machine minds]] · [[Continent:mind|Mind]]
Revision 174 as rendered
Current methods, human feedback and constitutional training, work in practice and come with no guarantees. The problem grows harder as systems become more capable than the people overseeing them.