Vitalik Buterin Sounds Alarm: Superintelligent AI Could Arrive Sooner Than You Think

by

The inventor of Ethereum, Vitalik Buterin, has sounded warnings regarding potential dangers linked to highly intelligent artificial intelligence and the importance of having robust safeguards in place.

At this point in time, as advancements in artificial intelligence are accelerating at an unprecedented pace, there’s been a substantial increase in worries related to the safety aspect of AI, and these thoughts were recently expressed by Buterin.

Buterin’s AI Regulation Plan: Liability, Pause Buttons, and International Control

On January 5th, Vitalik Buterin shared a concept he calls ‘defensive acceleration (d/acc),’ which involves creating technology primarily for protection rather than causing harm. It’s worth noting that this isn’t the first time Buterin has addressed potential risks in Artificial Intelligence.

In 2023, Vitalik Buterin stated one potential catastrophic outcome of AI malfunctioning is almost unimaginably dire: it might lead to the end of humanity itself.

In 2023, Buterin presented his theories, and now he’s moving forward with them. As per Buterin, the advent of superintelligence might be only a few years away.

It appears that AGI and superintelligence could become a reality within six years. To prevent potential catastrophe or irreversible harm, we can’t just speed up the progress of what’s good; we also need to curb the development of what’s harmful,” Buterin suggested.

To reduce potential dangers from artificial intelligence, Buterin proposes developing autonomous AI structures that closely align with human judgment. This way, AI continues to serve as a useful tool for humans, thereby decreasing the likelihood of disastrous consequences.

Later on, Buterin discussed the possibility of military forces being the ones to handle an ‘AI catastrophe’ situation. The increasing use of AI by military forces is evident, as demonstrated in Ukraine and Gaza. Furthermore, he expressed his concern that any regulation implemented for AI would likely exclude the military sector, making them a potential danger.

In a subsequent discussion, the co-founder of Ethereum detailed his vision for controlling AI usage. He emphasized that the initial move towards mitigating risks related to AI should involve holding users accountable.

As an analyst, I’d rephrase that statement like this: “The relationship between the creation of a model and its eventual application can sometimes be murky, but it’s ultimately up to the user to determine how AI gets put into practice,” Buterin emphasized, underscoring the significant influence users hold.

If the current regulations for liabilities aren’t effective, a potential solution could involve introducing “temporary halt” controls, which would enable AI oversight to temporarily decelerate progress in areas where harmful advancements might occur.

We aim to temporarily cut global computational power by around 90-99% during a crucial timeframe of 1-2 years. This drastic reduction would give humanity additional time to ready ourselves for whatever challenges lie ahead.

He said the pause can be implemented by AI location verification and registration.

An alternative method is managing the physical hardware of AI, as Buterin suggests, this hardware could be enhanced with a specific chip for regulation purposes.

As a researcher, I’m involved in a project where the AI systems can operate effectively only after receiving three endorsements weekly from international organizations. It’s crucial to note that at least one of these organizations must not have any military ties.

Nevertheless, Buterin admitted that his strategies have holes and are only ‘temporary stopgaps.’

Read More

2025-01-06 09:09