Apple spends fortune on Shutterstock to train its AI with vast library of images and videos

[Coloroscopio]

Tech giant Apple has struck a significant deal with stock media leader Shutterstock to fuel its development of artificial intelligence. According to a Reuters report, Apple has licensed millions of images, videos, and audio files from Shutterstock's vast collection. While the company might be late to the AI game, it’s probably better late than never. This partnership highlights its growing focus on in-house AI development and the crucial role high-quality data plays in the process, while positioning it alongside other tech behemoths like Google and Meta.

 

The Cupertino-based company has allocated $25 million to $50 million to tap into Shutterstock’s extensive collection of images, videos, and audio content. These licensed assets will serve as the building blocks for training Apple’s AI models, enabling them to recognize patterns, learn from data, and make informed decisions.

 

While Apple has remained tight-lipped about the specifics of the agreement, the influx of diverse content from Shutterstock could primarily benefit Apple’s image recognition and editing functionalities on iPhones and other devices. By training its AI models on a massive dataset of labeled images and videos, Apple can significantly improve the accuracy and efficiency of tasks like object identification, scene understanding, and automatic image organization within your photo library.

 

The implications extend beyond just photos. Shutterstock’s library also boasts a rich collection of audio and video content. This could potentially enhance voice recognition features like Siri, leading to more accurate voice commands and dictation. Automatic captioning and transcription could also see improvements as Apple's AI learns to better interpret audio data.

 

The Bigger Picture

 

It’s worth noting that Apple isn’t the first tech company to leverage Shutterstock's resources. Major players like Google, Amazon, and Meta Platforms have reportedly already utilized Shutterstock’s data to train their own AI systems.

 

This partnership is just one piece of the puzzle for Apple. The company is also reportedly in talks with various publishers to secure access to text-based content for training its large language models. By combining these data sources, Apple aims to create a well-rounded AI foundation that can power a variety of features and functionalities across its device ecosystem.

 

The alliance also builds upon Apple’s recent advancements in AI, particularly the development of its on-device AI system called ReALM (Reference Resolution As Language Modeling). Unveiled earlier this year, ReALM focuses on improving conversation understanding through its unique approach to deciphering references within dialogue. This allows ReALM to better grasp the context of a conversation and respond accordingly, potentially leading to more natural and productive interactions with virtual assistants like Siri.

 

Rumors swirling suggest Apple might be venturing into hardware robotics as well. These personal home robots, still in the early stages of conception, are envisioned to be mobile and could potentially follow users around their homes.

 

So, how do you envision AI impacting your Apple devices in the future?

 

• https://www.inc.com/kit-eaton/apple-signs-deal-for-ai-training-data-from-image-service-shutterstock.html 
• https://www.nasdaq.com/articles/apple-strikes-deal-with-shutterstock-for-ai-image-training 
• https://www.pymnts.com/cpi-posts/apple-partners-with-shutterstock-for-ai-image-training/
• https://www.reuters.com/technology/inside-big-techs-underground-race-buy-ai-training-data-2024-04-05/

 

 

Image: Lester69 | Dreamstime.com