In the digital arena where tech giants are fighting a fierce battle over artificial intelligence, Apple seems to be the poor relation. Faced with rivals like Meta, OpenAI, Microsoft, Google… clues indicate that Apple would not remain inactive and is preparing its response: Ajax and Apple GPT.
Source: Jackie Alexander on Unsplash
In the race for artificial intelligence that is being played out between the digital giants, Apple is lagging behind. While Meta (Facebook) recently made its language processing model (LLM), called Llama 2, open source, Google has also unveiled its own tool, PaLM, which powers Google Bard. For its part, OpenAI, supported by Microsoft, is leading the development of GPT-4, the engine behind ChatGPT and Bing. But where is Apple in all of this?
With the advent of chatbots, technologies like Apple Siri, Amazon Alexa and Google Assistant now seem somewhat outdated. Apple’s AI flagship, the voice assistant Siri, has shown signs of stagnation in recent years, and the company has remained remarkably quiet on the subject.
According to information relayed by Bloomberg and journalist Mark Gurman, Apple specialist, the company is however working behind the scenes on AI tools that could compete with those of OpenAI, Google and others, but without clearly defined strategy to make them accessible to the general public.
Ajax and Apple GPT: Apple’s silent response
According to these same sources, Apple is working hard to develop next-generation generative AI features to catch up with OpenAI and Google. The company would thus have created a framework called Apaxintended to design LLMs, and would have deployed Apple GPT internally, with the aim of launching a consumer product in 2024.
However, this information should be taken with a grain of salt. Indeed, Apple CEO Tim Cook has been rather reserved about the arrival of these new AI services on the market. If the technology has potential, “ a number of problems remain to be solved“, he said during a conference call in May.
According to people familiar with the project, the Ajax framework would have been created in 2022 in order to unify the development of machine learning within Apple. AI-related enhancements have already been rolled out to Search, Siri, and Maps based on this system. Today, Ajax would be used to create large language models and serve as the basis for the internal ChatGPT styling tool. A chatbot application was even created on an experimental basis at the end of 2022 by a small team of engineers. Although its rollout was initially put on hold due to security concerns over generative AI, it has since reportedly expanded to more employees.
However, the tool would require special permission for access, and one major constraint remains: it cannot be used to develop customer-facing features. Nevertheless, Apple employees would use it to help design product prototypes, to summarize text, and to answer questions based on the data it was trained with.
Apple isn’t the only company taking this approach. Groups like Samsung have also developed their own internal ChatGPT-like tools, after concerns were raised about third-party services leaking sensitive data.
According to Apple employees, the company’s tool essentially replicates Bard, ChatGPT and Bing AI, and does not include any new features or technologies. The system is accessible as a web application and has a simplified design not intended for the general public. As such, Apple reportedly has no plans to bring it to consumers.