A blog article by Gary Marcus discussing the risks that come from poor control over large language model (LLM) AIs like ChatGPT. This article outlines the ways that LLMs could be used by criminals to gather personal information from online sources and to help with terrorist propaganda.  

Format: Blog post 

Estimated Reading Time: 7 minutes  

Resource Link:  

AI risk doesn't equal AGI risk 

Topics covered:  

• The odds of creating an AGI The risk posed by LLM like ChatGPT 
• How the real risk is control over AI 

How to use this resource:  

This resource is good for highlighting the risks posed by AI models like ChatGPT do not come from the models themselves but how humans might use the models for harmful and destructive means like criminal activity. It gives insight into the need to build a framework to control and restrict AI use.  

Source Creditability: Gary Marcus is a professor of psychology and neuroscience at New York University. He has researched and published studies on Artificial Intelligence.