Vitalik Buterin Talks About AI Development Risks and the Technology’s Interaction with Cryptocurrencies
- The Ethereum co-founder participated in the Ethereum x AI Confluence Day event, which the Incrypted team attended.
- During the panel discussion, he and the founder of Stability AI discussed the development of AI and its interaction with humans.
During the Ethereum x AI Confluence Day event, Ethereum co-founder Vitalik Buterin and Stability AI founder Emad Mostaque discussed various aspects of artificial intelligence (AI) development and shared insights into the application options for technologies based on it. The Incrypted team also visited the event and gathered the most interesting highlights of their conversation.
According to Buterin, artificial intelligence poses two main risks. One is the rapid development of technology to the level of superintelligence, which humanity is not prepared for. The second risk is that AI is capable of significantly over-empower a “centralized actor,” which would allow it to “lock the world” into a kind of totalitarian system.
“Both of these risks are very important to think about, and I think most people working on AI want generally want to not do,” he emphasized.
The Ethereum co-founder also highlighted the term “decentralization.” He noted that the word has a broad meaning. Some people perceive it in terms of projects from the blockchain industry, while others it understand differently. In terms of AI-based technologies, the word also has its own meaning, the programmer believes.
“Ultimately, ithe most decentralized AI is the thing that you’re running yourself,” Buterin stated.
As an example of a decentralized AI model, he mentioned one version of the large language model LLaMa 3, which is capable of running using an Nvidia 4070 graphics card. Buterin believes that we are it’s not about a global network with nodes around the world but rather a technology accessible to everyone.
“It’s about figuring out how do we actually create a version of this technology that empowers everyone, instead of empowering a very small group of people at best. And the philosophy of open source and software, the philosophy of openness and AI, the philosophy of decentralization and crypto and all the different ways it’s interpreting, should be viewed as different ways of achieving that underlying goal,” said the Ethereum co-founder.
Discussing AI applications, the panelists noted several areas in which the technology would be particularly useful. Emad Mostaque believes that some of the most promising areas are medicine, and education.
In turn, Buterin talked about “defensive AI” that could be used at the local level. He believes it could help fight fraud and warn people of potential risks.
“An AI capable of reading your messages could signal potential scams by saying, ‘this message looks like a scam; seven thousand people have already gotten something like this.’ Such an AI can work locally to protect your data and provide valuable insights,” Buterin noted.
The experts concluded that “defensive AI” models will play an important role in the future and need attention from specialists. At the same time, Buterin noted the importance of maintaining the security of personal data in such cases.
During the discussion, the Ethereum co-founder touched on the interaction between the cryptocurrency industry and artificial intelligence technologies. He stated that AI has been part of the Web3 industry for over a decade and cited arbitrage bots trading on AMM as an example.
Buterin considers prediction markets to be one of the interesting applications of AI in the cryptosphere. This technology will be able to counter “bad actors” and reduce the impact of repetitive patterns of reasoning that are often found in this segment, the programmer believes. AI will also create fairer and more efficient prediction markets, Buterin believes.
“So, you have this crypto mechanism, which is dome kind of very decentralized, very democratic voting scheme. Then you have cryptocurrency incentives, you have AI, that provides low cost speed. And we combine it all with each other. I think that one of those crypto plus AI use cases that I think is very underrated,” the Ethereum co-founder summarized.
Mostaque expressed the opinion that crypto could help the development of open-source AI.
“Bitcoin, for example, consumes 160 TW*h of energy per year — more than half of what all the data centers in the world use. We could use this energy for intelligent currency systems and public goods such as safer AI models,” the entrepreneur noted.
At the end of the discussion, Buterin summarized that our future depends on how we align AI incentives with ethical standards. If people focus on creating models and datasets that prioritize safety, there will be much less motivation to develop dangerous AI, the programmer said.
