You idiot. They were referring to Anthropic’s models explicitly, but it could be a downstream issue (it is also true of every company providing AI models). If the law were to ever hold them accountable, there would potentially be huge liability which would be uninsurable.