Part of the source code of the AI assistant Claude, accidentally leaked by the company Anthropic
The American company Anthropic confirmed on Tuesday that part of the internal source code of its popular programming assistant based on artificial intelligence, Claude Code, was accidentally disclosed, reports CNBC, quoted by News.ro.
“No sensitive customer data or credentials were involved or exposed. It was a software version packaging issue caused by human error, not a security breach. We are implementing measures to prevent this from happening again,” a company spokesperson said in a statement.
A source code leak is a blow to the startup because it could give software developers and Anthropic competitors clues about how the programming tool that has become wildly popular was built.
A post on social network X that included a link to the Anthropic code has garnered more than 21 million views since it was published at 4:23 a.m. ET on Tuesday.
The incident is the second major data mishandling case for Anthropic in less than a week. According to Fortune magazine, descriptions of the company's future artificial intelligence model and other documents were recently discovered in a publicly accessible data archive.
Anthropic was founded in 2021 by a group of former OpenAI executives and researchers and is best known for developing the Claude family of artificial intelligence models.
The company released Claude Code to the general public in May, and the tool helps software developers build functionality, fix bugs, and automate various programming tasks.
Claude Code has seen massive adoption over the past year, and the product's estimated recurring revenue topped $2.5 billion in February.
The success of the tool led companies such as OpenAI, Google and xAI to invest heavily in the development of competing products.
