Triggered GPT’s safety guardrails for the first time

Technology and Creation AI
,
1

GPT is acting so weird tonight—it feels like it’s simultaneously dumbed down and not dumbed down.
The safety guardrails on GPT were triggered for the first time, and I still have no idea what caused it.

As shown
As shown905×101 3.8 KB

I thought it was sites in the prompt (since OAI has them themselves), but no matter how I changed it, nothing worked.
As shown
As shown949×218 7.1 KB

About two weeks ago, I asked GPT to execute the exact same instructions, and this wasn’t triggered.
I have no idea what’s going on.
Even Grok doesn’t know
Even Grok doesn’t know1280×745 141 KB

By the way, it also became impossible to update tonight.

*This is the first time I’ve encountered this*
*This is the first time I’ve encountered this*442×215 6.91 KB

Could Astra be coming soon?

1 Like
2

It looks like it’s working now—could you try again?

1 Like
3

Sure enough, I passed without changing anything. :xhj24:
But I still can’t update.

1 Like
4

Looks like OAI is up to no good again.

5

That’s what I’m saying too.

6

I’ve also run into the safety filter suddenly triggering several times before. I spent ages tweaking the prompt and nothing worked, then woke up the next day and everything was fine—probably just OAI’s backend acting up temporarily? It is pretty strange that the exact same instructions worked two weeks ago but suddenly don’t now. Maybe try rephrasing “sites” to get around it? I also had the update issue last week; it fixed itself after two days. Maybe it really is just the usual instability before a new feature rolls out. Perhaps Astra really is coming, haha.

1 Like
7

It started working after I left it for a while—seems like OAI was acting up :ac33:

8

It was glitching—there were a lot of cases like this yesterday.