Spirit airlines' data becomes valuable asset for AI firms amid bankruptcy

Spirit Airlines, which has been grounded since May as it undergoes bankruptcy proceedings, may soon receive a multimillion-dollar payout from the sale of its corporate data. This data includes internal wikis, emails, source code, and spreadsheets that are valuable for training artificial intelligence. Google successfully bid $10 million for the data, though AI training provider Micro1 submitted a higher offer of $12.5 million after the formal auction had concluded. The data sale is being managed through the bankruptcy process, and an attorney handling the sale did not immediately respond to inquiries. Under the Google deal, any personally identifiable information will be processed by a third-party deidentification agent, and the dataset does not include customer lists. However, a Spirit flight attendant union has objected to the sale, arguing that employee privacy could still be compromised even after names are removed. A Google spokesperson stated that the company will not receive any personal information from the dataset. The bidding for Spirit’s data is part of a broader trend where AI firms are acquiring datasets from both shuttered and operating businesses to improve AI models for real-world business environments. AI training company Mercor, which had offered $7.5 million for the data, emphasized that corporate records are invaluable for training AI to handle realistic office tasks. Privacy advocates caution that even anonymized data could expose sensitive information about employees and customers. The sale reflects a wider trend of companies acquiring data from various sources, including Reddit’s forum archives and out-of-print books, to enhance AI models. Additionally, AI firms are paying experts in fields such as finance and poetry to help train and test AI systems. These data acquisitions highlight the growing importance of corporate records in AI development, despite concerns over privacy and data security.