Welcome to our

Cyber Security News Aggregator

.

Cyber Tzar

provide a

cyber security risk management

platform; including automated penetration tests and risk assesments culminating in a "cyber risk score" out of 1,000, just like a credit score.

Subliminal Learning in AIs

published on 2025-07-25 11:10:10 UTC by Bruce Schneier
Content:

Today’s freaky LLM behavior:

We study subliminal learning, a surprising phenomenon where language models learn traits from model-generated data that is semantically unrelated to those traits. For example, a “student” model learns to prefer owls when trained on sequences of numbers generated by a “teacher” model that prefers owls. This same phenomenon can transmit misalignment through data that appears completely benign. This effect only occurs when the teacher and student share the same base model.

Interesting security implications.

I am more convinced than ever that we need serious research into AI integrity if we are ever going to have trustworthy AI.

Article: Subliminal Learning in AIs - published 3 months ago.

https://www.schneier.com/blog/archives/2025/07/subliminal-learning-in-ais.html   
Published: 2025 07 25 11:10:10
Received: 2025 07 25 11:18:46
Feed: Schneier on Security
Source: Schneier on Security
Category: Cyber Security
Topic: Cyber Security
Views: 8

Custom HTML Block

Click to Open Code Editor