Tech giant Apple is set to collaborate with stock image company Shutterstock in a deal estimated to be worth between US$25 million to US$50 million.
Shutterstock's Chief Financial Officer, Jarrod Yahes, has previously revealed that the signing amounts with various tech companies are of that range. This follows the rumored attempts as reported by Reuters from tech giants such as Apple to license the treasure trove of visual data that is Photobucket's 13 billion photos and videos to train its proprietary AI.
The partnership aims to acquire images and videos for training AI models. Analysts suggest that Apple's entry into the AI data training arena will intensify competition for acquiring diverse content.
The unquenchable thirst for more data
According to VentureBeat, sources reveal that Apple, Meta, Google, and Amazon have all partnered with Shutterstock to access vast libraries of images, videos, and music files.
Established in 2003, Shutterstock has emerged as a premier global hub for top-notch stock images, videos, and music files. Their website proudly boasts ownership of over 300 million images, with a staggering 200,000 more added daily by a dedicated community of over 1 million contributors.
With its expansive collection and user-friendly interface, Shutterstock has solidified its status as the ultimate destination for creative professionals, businesses, and individuals in need of visual content for diverse projects and purposes. Meanwhile, Photobucket, another player born in 2003, once reigned as the leading image-hosting site worldwide, as reported by Reuters.
At its zenith in the early 2000s, Photobucket boasted 70 million users and commanded nearly half of the U.S. online photo market. However, its user base has dwindled to a mere 2 million active users since then.
The growing demand for AI training data has spurred a thriving market, with AI data companies such as Defined.ai witnessing increasing willingness from content providers to pay for images and videos. The typical rates range from US$1 to US$2 per image, US$2 to US$4 per short video, and US$100 to US$300 per long video.
With AI advancements in recent years, the competition for data acquisition has become fiercer. Tech giants such as OpenAI, Google, Meta, and Microsoft have been utilizing large amounts of online data to train their AI models without acquiring approval or compensating content creators, including copyrighted news, books, and social media posts.
Regulation or not?
Some advocate for establishing licensing systems where AI companies compensate content owners to gain the right to use training data. The National Association of Broadcasters, News Media Alliance, and Condé Nast executives have expressed support for mandatory licensing.
However, OpenAI and some experts argue that universal licensing of all training data is impractical and that mandatory licensing could concentrate power in large tech companies and burden AI startups.
The collaboration between Apple and Shutterstock underscores the critical role of data in developing cutting-edge AI systems. As companies such as Apple, Google, and Meta race to build the most advanced AI models, access to extensive and diverse datasets has become a crucial dividing line.
By licensing millions of images from sites such as Shutterstock and Photobucket, Apple aims to broaden the scope of its AI capabilities, spanning from computer vision and image generation to virtual assistants and augmented reality (AR).
Moreover, the willingness of tech giants to pay millions for AI training data highlights the significant economic potential of this technology. Companies such as Apple are heavily investing in AI development to secure a prominent position in this thriving market.
According to Business Research Insights, the rapidly growing AI data market is valued at around US$25 billion and is projected to grow to approximately US$30 billion over the next decade. This underscores the high stakes for tech giants vying for AI dominance, while the long-term impact on user privacy and data rights remains to be seen.
