Concrete Problems in AI Safety, Revisited

TL;DR

This paper revisits core AI safety problems, emphasizing the need for a broader socio-technical perspective to better understand and address failures in AI systems during real-world deployment.

Contribution

It analyzes real-world AI safety incidents and advocates for an expanded socio-technical framework to improve understanding of AI system failures and successes.

Findings

Current vocabulary captures some issues but is insufficient.

A socio-technical framing is necessary for comprehensive understanding.

Real-world cases highlight the complexity of AI safety challenges.

Abstract

As AI systems proliferate in society, the AI community is increasingly preoccupied with the concept of AI Safety, namely the prevention of failures due to accidents that arise from an unanticipated departure of a system's behavior from designer intent in AI deployment. We demonstrate through an analysis of real world cases of such incidents that although current vocabulary captures a range of the encountered issues of AI deployment, an expanded socio-technical framing will be required for a more complete understanding of how AI systems and implemented safety mechanisms fail and succeed in real life.