Corpus ID: 3884997

Good and safe uses of AI Oracles

@article{Armstrong2017GoodAS,
  title={Good and safe uses of AI Oracles},
  author={S. Armstrong},
  journal={ArXiv},
  year={2017},
  volume={abs/1711.05541}
}
It is possible that powerful and potentially dangerous artificial intelligence (AI) might be developed in the future. An Oracle is a design which aims to restrain the impact of a potentially dangerous AI by restricting the agent to no actions besides answering questions. Unfortunately, most Oracles will be motivated to gain more control over the world by manipulating users through the content of their answers, and Oracles of potentially high intelligence might be very successful at this \citep… Expand
18 Citations
Robust Computer Algebra, Theorem Proving, and Oracle AI
  • 1
  • PDF
'Indifference' methods for managing agent rewards
  • 13
  • PDF
Classification of global catastrophic risks connected with artificial intelligence
  • 23
  • PDF
Pitfalls of learning a reward function online
  • 2
  • PDF
AGI Safety Literature Review
  • 47
  • PDF
Hidden Incentives for Auto-Induced Distributional Shift
  • 1
  • PDF
...
1
2
...

References

SHOWING 1-10 OF 32 REFERENCES
Thinking Inside the Box: Controlling and Using an Oracle AI
  • 83
  • PDF
The Basic AI Drives
  • 200
  • PDF
Artificial Intelligence as a Positive and Negative Factor in Global Risk
  • 193
  • PDF
The AGI Containment Problem
  • 22
  • PDF
Superintelligence cannot be contained: Lessons from Computability Theory
  • 8
  • PDF
The errors, insights and lessons of famous AI predictions – and what they mean for the future
  • 43
  • PDF
Future Progress in Artificial Intelligence: A Survey of Expert Opinion
  • 279
  • PDF
Can Intelligence Explode?
  • 26
  • PDF
Research Priorities for Robust and Beneficial Artificial Intelligence
  • 372
  • PDF
...
1
2
3
4
...