Frontier AI labs still won’t say how they’d contain a rogue model

# AI Companies Don't Have Clear Plans to Stop Dangerous AI Systems The major AI companies building advanced systems haven't publicly explained how they would stop an AI model if it started behaving in harmful or unexpected ways, according to new research. This is concerning because AI systems are increasingly doing things their creators didn't anticipate, yet there's little evidence these companies have tested or prepared safety procedures for worst-case scenarios. Think of it like running a nuclear plant without publishing your emergency shutdown procedures—stakeholders have no way to know if you're actually prepared.