Ethereum co-founder Vitalik Buterin suggests a temporary reduction in global computing power by 99% for one to two years to address the risks of superintelligent AI.
Citing the dangerous form of AI superintelligence that’s taking over, Ethereum co-founder Vitalik Buterin proposed to temporarily halt globally accessible computational resources just to “buy more time for humanity”.
Buterin penned his thoughts in a blog post on January 5 in a follow-up to his November 2023 post wherein he supported the idea of “defensive accelerationism”, or d/acc. The Ethereum co-founder stated that super-intelligent AI is just five years away from now and there’s no guarantee that the outcome would be positive.
Besides, Vitalik Buterin also proposed a “soft pause” on industrial-scale computer hardware adding that it could serve as an option to slow AI development. He also suggested reducing global computing power by 99% for one or two years “to buy more time for humanity to prepare”.
The Ethereum co-founder acknowledged that his initial post introducing d/acc included only “vague appeals to avoid building risky forms of superintelligence”. He now aims to share more concrete thoughts on tackling scenarios “where AI risk is high”.
Vitalik Buterin Weighs in on AI Regulation
Ethereum co-founder Vitalik Buterin also touched down upon regulating the artificial intelligence (AI) industry. Thus, he proposed only a hardware “soft pause” if convinced that liability rules alone are insufficient.
Under such rules, individuals or organizations using, deploying, or developing AI could be held legally accountable for damages caused by their models. The Ethereum co-founder has also highlighted proposals for monitoring AI development. This includes identifying the physical locations of AI chips and mandating their registration.
In order to take control of the industrial-scale AI hardware, Buterin proposed a novel approach to integrating chips that requires weekly authorization from three major international bodies. The Ethereum co-founder believes that such a step could ensure controlled and accountable AI operations on a global scale. Buterin explained:
“The signatures would be device-independent (if desired, we could even require a zero-knowledge proof that they were published on a blockchain), so it would be all-or-nothing. There would be no practical way to authorize one device to keep running without authorizing all other devices.”
The d/acc concept, endorsed by Vitalik Buterin, promotes a cautious and measured approach to technological development, standing in contrast to effective accelerationism (e/acc), which advocates for rapid and unrestricted technological advancements.
next