Debate Over Apocalypse Probability in the Artificial Intelligence Sector
Apocalypse scenarios and percentage rates shared by artificial intelligence experts are increasing criticisms within the community that they lack a scientific basis and are merely marketing tactics.
The figures given by figures in the artificial intelligence sector ranging from 5 percent to 80 percent for risks that could bring about the end of humanity have ignited debates over whether these predictions have any scientific basis.
Artificial Intelligence Apocalypse Probability
Many figures from the artificial intelligence sector have shared various percentages regarding the possibility that the technology they develop could bring about the end of humanity, and this concept has been termed p(doom), meaning the probability of doom in Silicon Valley.
Evan Hubinger, alignment research lead at Anthropic, stated that there is a more than 10 percent chance that artificial intelligence will kill all humans within the next decade.
Different Percentages from Experts
Anthropic co-founder and CEO Dario Amodei stated that his personal apocalypse probability ranges between 10 and 25 percent.
Former Twitch CEO and brief OpenAI head Emmett Shear estimated this rate between 5 and 50 percent, while former OpenAI researcher Daniel Kokotajlo set it at 70 percent.
Criticisms and Doubts
Critics point out that no one has a real mathematical formula to calculate the artificial intelligence apocalypse probability and that subjective predictions add confusion to public debates.
Some skeptics argue that these warnings are public relations stunts and marketing moves designed to generate awe prior to public offerings.
Researchers' Defense
Georgetown computer scientist Cal Newport criticized this apocalypse fear-mongering strategy in his New York Times column.
In response, Anthropic artificial intelligence safety researcher Drake Thomas expressed that researchers are genuinely afraid rather than engaging in marketing.