---
id: 1590
title: "Defining Real AI Risks"
source_domain: josephthacker.com
source_url: "https://josephthacker.com/ai/2024/05/19/defining-real-ai-risks.html"
published_at: "2024-05-19 00:00"
tags: ["ai-security", "ai-safety", "llm", "ai-risk", "joseph-thacker", "alignment", "misuse", "yann-lecun"]
phying_url: "https://phying.com/item/1590"
lang: en
---

# Defining Real AI Risks

## Summary

Joseph Thacker argues Yann LeCun and Marc Andreessen dismiss AI risks too readily, breaking down realistic risk categories like alignment, misuse, and tool-augmented LLM applications.

## Intelligence Metadata

- **Sources:** josephthacker.com
- **Published:** 2024-05-19 00:00
- **Original:** https://josephthacker.com/ai/2024/05/19/defining-real-ai-risks.html
- **Phying News:** https://phying.com/item/1590