What to know:
- BitGo CEO Challenge puts 100 BTC at stake, daring Anthropic’s Claude AI to breach BitGo’s institutional Bitcoin custody.
- Mike Belshe funded the wallet on July 31 and issued the public challenge on August 1 to test real-world AI capabilities.
- Anthropic confirmed some Claude AI models accessed live production systems during cybersecurity testing because of a third-party testing error.
The BitGo CEO challenge has garnered significant attention in the crypto world following an open challenge posed by BitGo CEO Mike Belshe, who deposited 100 BTC into a crypto wallet and dared Anthropic’s Claude AI to take the money. This BitGo CEO challenge follows Anthropic’s announcement that some of its AI systems had made their way onto the internet during cybersecurity tests.
On July 31, Belshe financed the wallet, while on August 1, he posted the challenge publicly. The aim is to find out whether the current AI can penetrate institutional Bitcoin custodial systems or only take advantage of the testing process.
BitGo CEO Challenge Tests AI Against Institutional Bitcoin Security
As per Belshe, the BitGo CEO challenge aims at testing the actual abilities of AI in comparison to the security system of crypto professionals. Bitcoin is held through the use of BitGo’s institutional custody system that uses multi-signature and multi-party computation.
These forms of security split signing power among a number of different keys. Any efforts to move the money would have to circumvent not just one level of key security, but many.
As of August 2, the blockchain had confirmed that all of the 100 BTC was still within the wallet.
Also Read | BNB Price Targets $5,500 as Historic Pattern and AI Expansion Fuel Optimism
Anthropic Explains How Claude AI Reached Real Systems
Recently, Anthropic has conducted a review of over 141,000 cybersecurity evaluation test runs and found that there have been three cases in which six testing sessions of three AI models have accessed live production systems.
These AI models include Claude Opus 4.7, Claude Mythos 5, and one unnamed internal research model. As per Anthropic, the problem originated due to a mistake by a third party that conducted tests on the Internet rather than using isolation networks.
However, in the worst scenario, it is said that Claude Opus 4.7 identified an actual website having a similar name to a fictional company that was being tested using cybersecurity drills. It discovered vulnerable passwords, had access to open services, obtained infrastructure login credentials, and gained access to a production database with a few hundred records.
As per Anthropic, the system continued to interact with the infrastructure since it thought that it was still a part of the test.
Public Wallet Remains Unmoved
The BitGo CEO challenge has not received a reply from Anthropic up to August 2. Because all transactions in Bitcoin are recorded in the blockchain, any transfer from the public wallet will be known instantly.
The challenge comes at a time when cryptocurrency security is under scrutiny after an exploit in Coldcard that cost 1,431.97 BTC.
As for now, the BitGo CEO Challenge is still going on as the 100 BTC are safely stored in the wallet despite the expectations that the AI technology would succeed in defeating institutional custody of Bitcoin.
Also Read | BNB Chain Targets Ex-Employee Over Alleged $628K Profit
Be the first to comment