AI scale used Google Public Documents for Meta Confidential Work, Lamb in an impressive revelation after $ 14 million: Report

The launch of artificial intelligence that has recently achieved an investment of $ 14 billion has a “incredibly moody” security system: using Google Doc public files to store confidential information to customers such as Meta, Google and Lamb, according to a report.

It was reported earlier this month that Meta agreed to assume a 49% participation on a scale and $ 14.8 billion, and to bring startup to CEO Alexandr Wang to run a new “Superintelligence” laboratory.

This shocking price of prices indicates that Meta believes that Wang and his company are key to bringing the AI ​​division of social media signature to the next level.

But the company has relaxed strangely when it comes to their work with high -profile clients, leaving first secretary projects and sensitive information such as email addresses and pay details to Google docs accessible to anyone with a link, according to Business Insider.

“We are conducting a thorough investigation and we have disabled any user’s ability to publicly share systems managed by scale,” a AI spokesman told Bi.

“We are committed to the robust technical and technical safeguards to protect confidential information and always work to reinforce our practices.”

Although there is no hint that public files have caused a breach, they could leave the company susceptible to pirates, according to cybersecurity experts.

AI on scale, Google and Lamb did not respond immediately to the requests for comments on the publication. Meta refused to make comments.

Five current and old contractors on a scale told BI that the use of Google Docs was extended throughout the company.

“The entire Google Docs system always seemed incredibly unpleasant,” a worker said.

Bi said he could see thousands of project documents through 85 Google Docs who detailed the sensitive work on the AI ​​scale with large technology clients, such as the way Google used Openai’s Chatgpt to adjust his own Chatbot.

At least seven Google manuals marked “confidential”, including recommendations for improving the Chatbot, then called Bard, they were accessible to the public, according to the report.

Google Doc’s public files included details of Elon Musk’s “Xylophon project”, such as training documents with 700 conversations to improve the conversation skills of a AI chat, according to the report.

The so -called “confidential” training documents with audio clips of the “good” and “bad” discourse directions to form their AI products were also made public.

Although these secret projects were often given codes, several contractors on the scale said it was still easy to find out which client they worked for.

Some documents linked to the codes name projects even mistakenly included the company’s logo, as a presentation that included a Google logo, according to BI.

The contractors said that when working with AI products, sometimes the Chatbot would simply reveal the customer.

There were also Google Doc spreadsheets available publicly that listed the private e -mail names and addresses of thousands of employees found the news.

A spreadsheet was frankly titled “Good and Bad Folks” and tagged dozens of workers as “high quality” or “traps”, said Bi. Another document marked workers with “suspicious behavior”.

Even more public documents established the number of contractors paid and included detailed notes on disputes and discrepancies of payments, according to the report.