Apple and OpenAI Embrace AI Trends: Securing Media Rights to Strengthen Language Models

2024-01-09

Major artificial intelligence companies have invested billions of dollars in generative AI, ensuring that the investment is substantial. In addition to focusing on the development of AI, newsrooms have also become a significant part of the AI industry.





Tech companies such as OpenAI, Google, and even Apple, which develop these AI models, are eager to sign contracts with media giants to exclusively obtain copyright information for training their language models.


Recently, OpenAI, the manufacturer of ChatGPT, made headlines by partnering with Axel Springer, the parent company of Politico and Business Insider. Information also reported that OpenAI is willing to pay between 1 and 5 million dollars annually to obtain the rights to use copyrighted news articles for training its AI models.


Negotiations between Apple and news publishers seem to be an attempt to catch up with the trend of generative artificial intelligence. However, some people disagree because the trillion-dollar company does not rely on AI as a marketing gimmick. Apple has been quietly conducting AI research and has not received as much media attention as other companies.


In addition, secrecy is deeply ingrained in Apple's DNA and culture. It is famous for keeping tight-lipped about product announcements and releases.


For example, in October 2023, Apple collaborated with Columbia University to release an open-source multimodal language model called Ferret. Initially, this paper received little attention, but a few weeks later, there was more discussion as the community became more interested in the potential of on-device LLM to power small devices.


Apple also published two research papers on 3D avatars and efficient language model inference techniques, which could enable more immersive visual experiences and allow complex AI systems to run on devices such as the iPhone and iPad.


In the first paper, the research team introduced HUGS (Human Gaussian Splats), a tool that converts videos captured from a single camera into 3D avatars. In the second paper, they found a way to use large language models like GPT-4 on regular devices without consuming too much memory.


The news of Apple's entry into open-source and on-device ML development came during a period when the company was focusing on hardware capabilities. The designer of GPT, OpenAI, may be working on designing a smartphone with renowned iPhone designer Jony Ive.


Apple's virtual assistant Siri has remained relatively unchanged since its launch, but given its existing capabilities, we can expect it to be rebuilt with generative AI. Apple CEO Tim Cook has stated that Apple's AI-related work is "in progress" but has not provided further details publicly.


While Apple is busy with media companies, they remain silent on how to apply generative AI in the news industry. Apple has already attracted a large audience of news enthusiasts to their devices. Some news executives are optimistic that Apple's actions could turn them into true partners. There are also some insiders who are excited about the long-term potential of this deal.