Apple partners with Shutterstock to train its AI model

Apple has teamed up with Shutterstock, a library of stock photos and videos, to use its data for training its artificial intelligence (AI) models. This partnership indicates that Apple is moving towards creating its own foundation model. While there were rumors about integrating Google’s Gemini AI and Baidu’s Ernie Bot into the iPhone 16 series for AI features, it appears that Apple’s focus is on developing its own AI model in the long run.

Apple has partnered with Shutterstock

Sources familiar with the matter cited by Reuters revealed that Apple has struck a deal with Shutterstock to access its licensed data, including images, videos and music, for AI training. Other tech giants such as Meta, Google and Amazon have reportedly signed similar agreements. Jarrod Yahes, Shutterstock’s CFO, revealed that these deals initially ranged from $25 million to $50 million each, with most companies then scaling up in size.

In addition to the Shutterstock deal, Apple has also inked deals with Seattle-based AI data company Defined.ai, which licenses data to tech companies to train their AI models. According to Daniela Braga, CEO of Defined.ai, typical prices for licensing data are around $1-$2 per image, $2-$4 per short-form video, and $100-$300 per hour for longer films. text is priced at $0.001 per word.

These developments point to two intriguing possibilities. First, the inclusion of Gemini AI and Ernie Bot in the iPhone 16 lineup may serve as temporary solutions until Apple finalizes its internal AI models. Therefore, while the original models have outsourced AI features, they will likely be replaced when Apple officially unveils its own AI technology. Second, the partnership with Shutterstock shows that Apple’s AI efforts extend beyond text capabilities to include images, music, and video.

Although the exact purpose behind the development of these AI models remains unclear, it is reasonable to conclude that they are multimodal and may include aspects of computer vision and image and video generation. In particular, Apple Research has published two separate papers describing AI models currently under development.