The race to harness artificial intelligence is escalating into a fierce battleground for content. Microsoft chief executive Satya Nadella testified Monday in a landmark U.S. trial against its rival Google, saying tech giants are competing for vast troves of the material needed to train artificial intelligence. He complained that Google, which dominates the online search market, was locking up content with expensive and exclusive deals with publishers. “It’s like the early days of distribution deals,” he said. The tech industry is racing to build a library of content to power large language models used by AI systems to process human speech and understand text. Those models, in turn, are used to answer queries from users.
But while computing power and data are vital elements in AI development, access to extensive content is critical for training sophisticated language models, Nadella said. He warned that exclusive deals could hinder fair competition, particularly as the technology reaches broader applications in healthcare and self-driving cars.
While he refrained from explicitly naming Google, Nadella pointed to its deals with publishers as a particular problem for competitors in the race to develop AI. The Justice Department’s antitrust case against Google, which has 90% of the online search market, hinges on the argument that the tech giant is abusing its dominance by paying billions of dollars each year to smartphone makers and wireless carriers to keep its Google Search app as the default option on their devices. This allegedly gives the company a dominant position in the search engine market, leading to dominance in the advertising business.
Nadella, who took over Microsoft in 2014, said the company focused on building “intelligent” products and helping customers get the most out of its AI capabilities. He also praised employees’ work in a team tasked with identifying risks posed by the company’s AI tools. The team was dissolved, though, as the rest of the company became more focused on shipping AI products faster than its rivals.
Employees who spoke to Reuters on condition of anonymity said they were disappointed the group was cut, especially since it was working to ensure that responsible AI principles are incorporated into the design of all Microsoft’s AI products. For instance, the team designed a role-playing game to help designers envision possible harms from their products and discuss them during product development.
The Justice Department’s antitrust case against Microsoft 1998, which led to a settlement that changed the company’s business practices, set a high bar for tech giants to take responsibility for their products. However, some experts have questioned whether companies such as Apple Inc., Facebook Inc., and Google are taking that burden seriously, especially after several public and internal incidents in recent years.
