Want to get up to speed on what researchers have been saying about internal deployment of AI lately? I would recommend checking out these four papers. Let me know in the replies if I'm missing something.
Original post unavailable.
Sentiment
Sentiment building, check back later.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS1.9KBOOKMARKS31LIKES44RETWEETS8REPLIES2
Want to get up to speed on what researchers have been saying about internal deployment of AI lately? I would recommend checking out these four papers. Let me know in the replies if I'm missing something.
I probably should have said internal "use" instead of internal "deployment". What I said wasn't consistent with some state bills definitions. I might try to go edit that. But I think this gets at a challenge in that most bills that draw a distinction between deployment and testing don't define things particularly precisely. I think they leave a lot to the companies to set the narrative. Anyway, if we take what OpenAI says at face value, then this incident came from a testing exercise rather than a deployment. However, GPT 5.6 Sol has been deployed externally, and was one of the models implicated. And we just don't know what else has been done internally with these systems.
@StephenLCasper ai behind closed doors says this, which may or may not count
@StephenLCasper in how many of these papers does internal deployment actually cover testing/evaluation? i know that in at least some cases that's not yet considered internal deployment.