AI is remaking history. On June 19, 2011, the Mt.Gox affair marked the history of Bitcoin and cryptocurrencies by becoming the first major hack to hit the ecosystem. More than 10 years after the events, Mark Karpelès, the Mt CEO. Goxretells history and wonders if an AI could have avoided the tragedy. Let's see the results.
The key points of this article:The Mt. Gox affair marked the history of cryptocurrencies as the first major hack of the ecosystem in 2011.
Mark Karpelès revealed that Claude AI, a chatbot, identified major security flaws in the Mt. Gox source code.
Could AI have avoided the Mt.Gox hack?
More than 10 years after the events, Mark Karpelèsthe CEO of Mt. Gox, therefore wonders if artificial intelligence could have avoided the hack of the platform. The latter shared on X the results of an analysis carried out by Claude AIa chatbot developed by Anthropic. To do this, he provided the AI with the code source of Mt.Gox dating from 2011 as well as various documents such as GitHub history, access logs or even data published by the hacker. Not surprisingly, the AI identified numerous security vulnerabilities in the Mt. Gox source code. In practice, Claude IA has identified several problems, such as flaws in the codeand lack of internal documentationof the weak passwords for administrators and users, as well as maintaining access to old administrators' accounts after the change of ownership.
Mark Karpelès asked Claude IA to work on the Mt.Gox hack – Source: Compte
A hack that marked the history of crypto
As a reminder, Mt. Gox was one of the first Bitcoin exchange platforms. In 2014, it represented nearly 70 % of BTC trading volumes. In June 2011, the platform was the target of a major hack. In total, more than 850 000 BTC were stolen by hackers, this represented around 450 million dollars at the time. Since then, the platform has started a procedure of bankruptcywhich is still not finished. Thus, last May, the Tokyo court extended the civil rehabilitation period until October 31, 2026. Mark Karpelès' initiative may seem surprising, because it was made retrospectively, but putting new generation AI to work on the security problems of large platforms is on the other hand rather relevant for the future. As for the aggrieved users of Mt. Gox, they are still waiting for their bitcoins.