A new video discussing big tech’s race towards artificial super intelligence has proven popular, and has some viewers worried.
Kurzgesagt, a science and education channel with over 20 million subscribers, has previously discussed topics such as astrobiology, factory farming, and existential risks like nuclear war. This week, their attention has turned to the prospect of artificial super intelligence (ASI) and its potentially catastrophic consequences.
Titled A.I. ‐ Humanity’s Final Invention?, the YouTube video has garnered over 2 million views since it was published. You can watch it here.
The video kicks off with a brief history of evolution, and explains how general intelligence, for most organisms, was simply too expensive. Narrow intelligences optimised to perform specific tasks make up the vast majority of the timeline of natural selection. But after a while, mammals came along with bigger brains built for more complex tasks, and eventually, about 2 million years ago, the first human was born.
Humans have radically transformed the landscape of the Earth in a miniscule amount of time (on an evolutionary scale), have sent burning towers of metal to other worlds, and have subjected trillions of other animals to the horrors of factory farming and animal testing. The one thing that has allowed us to dominate, that gives us the upper hand over all other life on Earth, is our superior intelligence.
What then, asks the video, would happen if we were to be knocked off top spot by a new species of our own creation?
Kurzgesagt consulted various experts during production, including AI safety educator Rob Miles and former member of the now-defunct ‘superalignment’ team at OpenAI, Daniel Kokotajlo, who quit after becoming disillusioned with the company’s reckless attitude towards safety.
Read more: OpenAI is the ‘Titanic of AI’, claims former safety researcher
Some viewers criticised Kurzgesagt for ‘selling out’, a reference to the funding they’ve received from Open Philanthropy (an effective altruist grant-making and research organisation) to create videos on topics related to ‘improving the long-term future’. It’s unclear why a charitable organisation using their resources to attempt to raise awareness around an important issue is cause for concern, or if the detractors had any substantive critique to make of Kurzgesagt’s video.
Max Tegmark, president of the Future of Life Institute, said the video is ‘the most important’ of 2024.
The most important video of 2024: if we build a new smarter-than-human species that doesn’t care about humans, it will be the dumbest thing we ever do.
Tegmark has claimed that big tech companies are intentionally distracting the world from what he sees as the very real existential threat of artificial intelligence. An open letter from Tegmark’s institute asking for a 6-month pause on the training of AI systems more powerful than GPT-4 was signed by AI researchers such as Yoshua Bengio and Stuart Russell, along with tech entrepreneurs Steve Wozniak and Elon Musk. The letter made waves in the AI world, receiving a total of 30,000 signatures, and inspiring the formation of grassroots movement PauseAI.
Kurzgesagt’s video also discusses the black-box nature of frontier AI models. Our lack of ability to understand their inner mechanisms is a huge problem for researchers attempting to align future systems with human values. This basic fact about the technology, however, has been denied by lobbyists in their attempts to stifle regulatory progress in the US.
Read more: AI lobbyists are blatantly lying to politicians
The video ends with the following quote:
The only thing we know for sure is that today, right now, many of the largest and richest companies in the world are racing to create ever more powerful AIs. Whatever our future is, we are running towards it.
