Apple uses data purchased from Shutterstock to train AI models: report

Apple has reportedly struck a deal with stock photo and video library Shutterstock to use the data to train its artificial intelligence (AI) models. The report also confirms speculation that the tech giant may be in the advanced stages of developing its in-house base model. Although previous reports mentioned that the iPhone 16 series may be equipped with AI capabilities from Google’s Gemini AI and Baidu’s Ernie Bot (China), Apple’s long-term plan seems to be to integrate native AI models.

in a ReportCiting unnamed people familiar with the matter, Reuters revealed that Apple has reached an agreement with Shutterstock to obtain its licensed data (images, videos and music) to train its artificial intelligence model. Apple isn’t alone, however, as Meta, Google and Amazon have reportedly struck similar deals. Shutterstock Chief Financial Officer Jarrod Yahes told Reuters that initial deal sizes for technology companies ranged from $25 million (roughly Rs. 208 crore) to $50 million (roughly Rs. 416 crore) each. According to CFOs, most companies have increased deal size.

This isn’t the Cupertino-based tech giant’s only deal to source data for its artificial intelligence models. It also struck a deal with Seattle-based Defined.ai, an AI data company, to license data to tech companies that train its AI models, according to the report. Defined.ai CEO Daniela Braga told Reuters that “companies are typically willing to pay $1 to $2 per image, $2 to $4 per short clip, and $100 to $300 per longer video. Text The market price is $0.001 per word.”

If this information is to be believed, it points to some interesting possibilities. First, the company’s rumored inclusion of Gemini AI and Ernie Bot may just be a stopgap measure until it can fine-tune its own AI models. This means that even if the iPhone 16 series is equipped with outsourced AI functions, they will eventually be replaced when Apple officially announces its own AI. Second, the Shutterstock deal highlights Apple’s focus not just on text-only functionality, but also on images, music, and video.

While it cannot be said whether it is just for developing computer vision or letting the models generate images and videos, it is safe to assume that the AI ​​models being developed are multi-modal. Notably, Apple Research published two separate papers on a series of artificial intelligence models it is working on.