Simon Goldstein
Simon Goldstein is an Associate Professor in philosophy at Australian Catholic University. He is interested in researching the ethics and safety of 'language agents', a new type of AI agent that implements a belief/desire folk psychology to accomplish goals with reasoning assistance from LLMs.
Research:
Shutdown-Seeking AI
Aggregating Utilities for Corrigible AI
Language Agents Reduce the Risk of Existential Catastrophe (Co-Authored with Cameron Kirk-Giannini, published in AI & Society)
The Polarity Problem (Co-Authored with Cameron Kirk-Giannini)
AI Deception: A Survey of Examples, Risks, and Potential Solutions (Co-Authored with Peter Park and other CAIS staff and affiliates).
Media:
Leading AI experts warn the technology could lead to the extinction of humanity (SBS News)
ABC Radio (1:22:30 - 1:30:15)
(ABC News Op-ed) Is it ethical to create generative agents? Is it safe? (Co-Authored with Cameron Kirk-Giannini)
Center for AI Safety resources :
