Buterin Sees AI as Cybersecurity Game-Changer

Vitalik Buterin, co-founder of Ethereum, has expressed his disagreement with the notion that improving artificial intelligence will make cybersecurity unwinnable. In a social media post, he argued that AI may actually give defenders the tools needed to build software that is significantly harder to exploit.

Buterin tied this belief to his own experience in cryptocurrency, stating that holding substantial amounts of cryptocurrency implies betting on the possibility of building sufficiently secure digital systems despite increasingly sophisticated attackers. He pointed out that if AI can prove complex mathematical theorems like Navier-Stokes and FLT, it should also be able to prove a program's security as a mathematical theorem.

The Ethereum co-founder emphasized the importance of formal verification, a method of mathematically demonstrating that software satisfies specific properties. However, he acknowledged that proving software is secure can be challenging due to complexities like the 'definition problem' and potential issues with corrupted databases, malicious libraries, or hardware-level information leaks.

Buterin also noted that while more capable AI systems could make finding vulnerabilities easier, they could also lower the cost of formally proving software behaves according to carefully constructed security specifications. Ultimately, both attackers and defenders will have access to AI tools, which could lead to a shift in favor of the defenders.