The concepts you propose are actually discussed at length in Nick Bostrom’s recent work, Superintelligence, although tripwires are only one of many security measures a safe design might employ. The book is a good introductory text on the various technical problems superintelligences (including AGI) carry with them, as well as some global-political overview.
I really recommend it, if you’re sufficiently interested.
The concepts you propose are actually discussed at length in Nick Bostrom’s recent work, Superintelligence, although tripwires are only one of many security measures a safe design might employ. The book is a good introductory text on the various technical problems superintelligences (including AGI) carry with them, as well as some global-political overview.
I really recommend it, if you’re sufficiently interested.