Gary Marcus

Rank 12 of 47

Score 92

Reposted by Gary Marcus

The title says it all:

🚨"Commercial LLM Agents Are Already Vulnerable to Simple Yet Dangerous Attacks." 🚨

Even worse: "All our attacks are trivial to implement and require no understanding of machine learning."

Companies: stop racing to release products that aren't ready!

2/13/2025, 3:42:53 PM

The statement raises concerns about the security vulnerabilities of commercial LLM agents, urging companies to prioritize safety over speed in product releases. It highlights the ease of implementing attacks, emphasizing the need for caution and responsibility.

Principle 1:
I will strive to do no harm with my words and actions.
The statement aims to prevent harm by warning about potential security risks, aligning with the principle of doing no harm. [+2]
Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
It promotes understanding and awareness of security issues, encouraging companies to act responsibly. [+2]
Principle 6:
I will use my influence for the betterment of society.
The statement uses its platform to advocate for better security practices, aiming to benefit society by preventing potential harm. [+2]