We Can't Really Affect AI Security

The blog post discusses the notion that the level of AI security is naturally determined by collective human behavior rather than individual efforts. It argues against the idea that individuals or groups can meaningfully alter this security baseline, suggesting that security will only increase in response to significant breaches that generate public demand for change. Instead, the author emphasizes that security dynamics operate independently, correlated more with human emotions and societal risks than proactive measures from security experts.