Google pays $10 million for 100 million Spirit Airlines emails and 500 million Teams chats to train AI
The proposed purchase covers employee emails, Microsoft Teams messages, spreadsheets, calendars, marketing material, aircraft operations, revenue management, and worker productivity data. Read Entire Article
Google has secured a $10 million auction win, obtaining 100 million emails and 500 million Microsoft Teams chats from the defunct Spirit Airlines. The tech giant plans to utilize this internal data to enhance its products and train AI models. However, the purchase is subject to federal bankruptcy court approval during a Wednesday hearing.
The data comprises employee communications, operational records, custom software, and various other files totaling approximately 17 million OneDrive items, over 20 million SharePoint assets, and 516 code repositories holding around 30 million lines of proprietary code. Google will also receive financial databases, IT support tickets, and customer-service workflows.
Unlike web-scrapped data, this collection allows for understanding decision-making processes, colleague coordination, mistake rectification, and system navigation, all of which could bolster AI agents designed for workplace tasks. The auction began at $5 million, with AI training startup Mercor entering as backup bidder at $7.5 million before Google's successful bid.
Google assures it won't acquire Spirit's customer or credit-card data, and a third party will anonymize the data before transfer, with Google covering the costs. Despite the anonymization process, removing personally identifiable information from decades of emails and chats may prove challenging, as sensitive details and contextual information can still reveal identities.
Spirit Airlines ceased operations this year after failing to emerge from its second Chapter 11 bankruptcy, and its remaining assets, including a wealth of data, are being sold by its estate attorneys. This transaction exemplifies the burgeoning market for failed startup data in the AI training materials race.
Written by urgent.news from TechSpot's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.