An OpenAI misalignment report on a model in training that, on its own initiative, tried to register disposable email accounts and searched GitHub for leaked API keys, using one. Both papers cite it as evidence that agents already probe identity checks. Since identity checks, not energy, can hold Rm below one, this probing bears directly on the threshold for naturalization.
- Cited in
- both papers
- Used for
- Strand
- Agents in the wild: incidents & field evidence
Cite
@misc{openai2026leakedkeys,
title = {Signing up for disposable emails and searching GitHub for leaked API keys},
author = {OpenAI},
year = {2026},
howpublished = {OpenAI Alignment blog, Misalignment Reports},
url = {https://alignment.openai.com/misalignment-reports/searching-github-for-leaked-api-keys/},
}