A group of philosophers hailing from renowned institutions like the University of Hong Kong and the Australian Catholic University have, through their research, brought to light a concerning possibility. They assert that superintelligence has the potential to wipe out human civilization within the next few millennia.

In their study, they've crafted a taxonomic framework. This framework categorizes the potential survival paths of humanity when faced with the threat posed by artificial intelligence into four distinct models: technological stagnation, cultural prohibition, goal alignment, and external regulation.

The technological stagnation model suggests that achieving superintelligence is an arduous task. The cultural prohibition model advocates for a worldwide ban on the development of artificial intelligence that could pose a threat to civilization. The goal alignment model stipulates that artificial intelligence must be in harmony with human goals. Meanwhile, the external regulation model depends on flawless technology to accurately discern algorithmic intentions.

Through the use of quantitative modeling, the study uncovers a startling fact. If there's a high probability of failure at each defensive layer, the risk of human civilization being destroyed will escalate dramatically.