In a revealing turn of events, Microsoft’s CEO Satya Nadella forwarded an email expressing deep concern over the capabilities of a rival company. The email, part of evidence presented in the Department of Justice’s antitrust case against Google, disclosed Microsoft’s apprehension and envy regarding Google’s advancements in AI technology. This apprehension led to Microsoft’s significant investment in OpenAI, a move driven by the urgent need to stay competitive.
The correspondence, a heavily redacted email thread from 2019 titled “Thoughts on OpenAI,” shed light on Microsoft’s worries about Google’s rapid progress in AI. CTO Kevin Scott, in the email forwarded by Nadella to CFO Amy Hood, voiced serious concerns about Google’s advancements, particularly in game-playing AI models like AlphaGo. Scott admitted to initially underestimating Google’s capabilities but grew increasingly alarmed as he witnessed the company’s advancements in natural language processing (NLP) models like BERT-large.
Microsoft’s attempts to replicate Google’s BERT-large model were hindered by infrastructure limitations, highlighting the gap between the two companies in terms of AI capabilities. Scott also praised Google’s achievements in Gmail auto-complete features, emphasizing Microsoft’s lag behind the competition in machine learning (ML) scale.
Prompted by Scott’s concerns, Nadella expressed his support for investing in OpenAI to address the company’s shortcomings. Microsoft subsequently made a billion-dollar investment in OpenAI in 2019, signaling a strategic move to bolster its AI capabilities. This investment has since grown substantially, reflecting Microsoft’s commitment to remaining competitive in the rapidly evolving AI landscape.
Meanwhile, OpenAI’s collaboration with leading publishers like the Financial Times has raised questions about copyright infringement. Several publications, including those owned by Alden Capital Group, have filed lawsuits against Microsoft and OpenAI, alleging unauthorized use of copyrighted materials to train AI models. These lawsuits underscore the growing tension between technology companies and publishers over the use of proprietary content in AI development.
