Apple In Talks With Leading News Outlets For A $50 Million AI Boost; Reports

Apple has embarked on discussions with major news and publishing entities to bolster its generative artificial intelligence systems. The California-based tech giant is reportedly seeking multiyear agreements, each valued at a minimum of $50 million, to obtain licenses for the archives of news articles. Sources familiar with the negotiations have revealed that Apple has engaged in talks with influential players such as Condé Nast, NBC News, and IAC, the owner of People, the Daily Beast, and Better Homes and Gardens.

The New York Times first reported on Friday that Apple's outreach to news entities signals a significant shift in the company's approach to leveraging external content for the advancement of its AI capabilities. The proposed agreements aim to provide Apple with access to a vast reservoir of news articles, potentially enhancing the depth and diversity of data used in its generative AI models.

However, not all news publishers have responded with equal enthusiasm to Apple's overtures. According to the report, some publishers approached by the tech giant exhibited a tepid response, reflecting a cautious stance within the industry about the implications of such collaborations.

Simultaneously, Apple is making strides in internal AI development, having reportedly created a service similar to ChatGPT. This internal tool is designed to aid employees in testing new features, summarizing text, and answering questions based on accumulated knowledge. Speculation has been rife about Apple's venture into AI, with July reports suggesting the company's development of a new AI model centred around a framework named Ajax. One of the potential applications mentioned was a ChatGPT-like interface, informally referred to as "Apple GPT."

Recent indications from an Apple research paper, initially uncovered by VentureBeat, add credence to these speculations. Titled "LLM in a Flash: Efficient Large Language Model Inference with Limited Memory," the paper addresses a critical challenge related to the on-device deployment of Large Language Models (LLMs) on devices with constrained DRAM capacity.

The primary author of the paper, Keivan Alizadeh, a Machine Learning Engineer at Apple, explained the approach, stating, "Our approach entails developing an inference cost model that aligns with the characteristics of flash memory, directing us to enhance optimization in two crucial aspects: minimizing the amount of data transferred from flash and reading data in larger, more cohesive segments."

This breakthrough in efficient large language model inference holds significant implications for Apple devices, including iPhones and iPads. The prospect of deploying Large Language Models directly on consumer devices opens new possibilities for improved performance and responsiveness, particularly in scenarios where memory resources are limited.

As Apple diversifies its AI initiatives, the engagement with news and publishing entities underscores the company's commitment to enriching its AI models with diverse and expansive datasets. The proposed multiyear agreements, if successfully negotiated, could pave the way for Apple to harness the collective knowledge and insights present in the vast archives of leading news outlets, ultimately enhancing the capabilities of its generative AI systems.

Industry experts are closely monitoring these developments, anticipating the potential impact on the landscape of AI applications and the role of major technology companies in shaping the future of artificial intelligence.

*Inputs from Reuters and Mint*