How much evidence for AGI do you think the public will get to see before it's locked away for good?

I remember back in 2019 OpenAI stated that they had a sweet spot where they didn't bother making anything more powerful than GPT-3, because they could release it to the public without any concerns, but the next leap in power was something they had to be much more careful with.<br><br>I don't think it's just the synthetic text that poses the risk, it's also the fact that your average person can now have access to a highly intelligent software agent that can assist them in virtually any way imaginable. It's like having a magical genie that guides them through every step of their lives.<br><br>I wonder if the government, and the megacorp giants will decide to directly collaborate and just lock away AGI entirely, and only use it with whatever purposes they deem fit.