Tech giants battling for content to train AI, Microsoft CEO says
Microsoft CEO Satya Nadella said that Google has moved aggressively to secure exclusive access to content from publishers with which to train AI models
Microsoft making its own AI chips is significant: R Ray Wang
Constellation Research founder R Ray Wang joins ‘Varney & Co.’ to discuss reports that Microsoft is pursuing a chip solution that would make them less dependent on Nvidia.
Big tech firms are locked in a competition to secure vast quantities of content and data to train artificial intelligence (AI) models, Microsoft CEO Satya Nadella said Monday.
Nadella was testifying in an antitrust trial the U.S. government has brought against Google, which is the first major tech-industry antitrust case in the U.S. since the government sued Microsoft in 1998. Nadella testified that Google has locked down content with costly exclusive deals and said that the race for content among tech firms "reminds me of the early phases of distribution deals."
The government’s antitrust case against Google contends that the company, which controls roughly 90% of the search market, has illegally paid $10 billion annually through distribution agreements to smartphone makers such as Apple and wireless carriers like AT&T to be the default search engine on their devices.
MICROSOFT TO ASSUME AI COPYRIGHT LIABILITY FOR COPILOT USERS
Satya Nadella, CEO of Microsoft Corp., said that big tech firms are locked in a race to secure access to content needed to train AI models. (Mark Kauzlarich/Bloomberg via Getty Images / Getty Images)
Nadella said that Microsoft is "happy to put in the dollars" to invest in servers needed to power AI models, but noted it has been "problematic" making deals with big content makers.
"When I am meeting with publishers now, they say Google’s going to write this check and it’s exclusive and you have to match it," Nadella said.
| Ticker | Security | Last | Change | Change % |
|---|---|---|---|---|
| MSFT | MICROSOFT CORP. | 423.46 | +0.47 | +0.11% |
| GOOGL | ALPHABET INC. | 168.95 | -0.28 | -0.17% |
Training the large language models that inform AI programs requires vast quantities of data, which is typically scraped from the internet — a practice that can raise liability for infringing intellectual property rights if that content is used without permission from the publisher.
WHAT IS ARTIFICIAL INTELLIGENCE (AI)?
Microsoft announced in September that it would assume copyright liability for users of its AI-powered Copilot. (Pavlo Gonchar/SOPA Images/LightRocket via Getty Images / Getty Images)
Several AI platforms, including OpenAI, which has received substantial investment from Microsoft, have been sued over the use of copyrighted materials in training AI models.
In September, Microsoft announced a Copilot Copyright Commitment under which it would assume any potential legal liability faced by users of its Microsoft 365 Copilot that can be used in Word, Excel, PowerPoint and other programs, as well as the company’s commercial Copilot services, Bing Chat Enterprise and GitHub Copilot. Users must abide by the Copilot's guardrails and filters to remain eligible for liability protection.
GET FOX BUSINESS ON THE GO BY CLICKING HERE
Microsoft said that it is "sensitive to the concerns of authors, and we believe that Microsoft rather than our customers should assume the responsibility to address them."
"It is critical for authors to retain control of their rights under copyright law and earn a healthy return on their creations," the company added in the September announcement. "And we should ensure that the content needed to train and ground AI models is not locked up in the hands of one or a few companies in ways that would stifle competition or innovation."
Reuters contributed to this report.