Without countermeasures, a scheming AI could escape.

A safety case (for deployment) is an argument that it is safe to deploy a particular AI system in a particular way.[1] For any existing LM-based system, the developer can make a great safety case by demonstrating that the system does not have dangerous capabilities.[2] No dangerous capabilities is a straightforward and solid kind of safety case. For future systems with dangerous capabilities, a safety case will require an argument that the system is safe to deploy despite those dangerous capabilities. In this post, I discuss the safety cases that the labs are currently planning to make, note that they ignore an important class of threats—namely threats from scheming AI escaping—and briefly discuss and recommend control-based safety cases.

1. Safety cases implicit in current safety plans: no dangerous capabilities and mitigations to prevent misuse

Four documents both (a) are endorsed by [...]

---

Outline:

(00:53) 1. Safety cases implicit in current safety plans: no dangerous capabilities and mitigations to prevent misuse

(03:32) 2. Scheming AI and escape during internal deployment

(07:41) 3. Control techniques and control-based safety cases

The original text contained 19 footnotes which were omitted from this narration.

---

First published:
June 3rd, 2024

Source:
https://www.lesswrong.com/posts/mmDJWDX5EXv6rymtM/companies-safety-plans-neglect-risks-from-scheming-ai

---

Narrated by TYPE III AUDIO.

“Companies’ safety plans neglect risks from scheming AI” by Zach Stein-Perlman

LessWrong (30+ Karma)