According to a recent report, Apple has developed its own large language model (LLM) AI framework and chatbot service, similar to those created by OpenAI (ChatGPT), Google (Bard), and Microsoft (ChatGPT-based Bing Chat). Internally known as “Ajax,” some engineers within the company have informally referred to it as “Apple GPT.” The framework is built on top of Google’s Jax machine learning framework and runs on Google Cloud.
Delays in its rollout due to security concerns
The project was initiated as an experiment around the end of 2022 and has been utilized by Apple employees for certain tasks, such as product prototyping. However, there have been delays in its rollout due to security concerns.
Currently, there is a company rule prohibiting its use for building consumer-facing products. Employees must go through an application and approval process to access the web-based app, which has a simple design not intended for public use.
While Apple has not yet finalized plans to introduce Ajax-driven features to consumers, executives John Giannandrea and Craig Federighi are leading teams working to figure out a strategy, with the intent to announce something in 2024. Apple’s AI strategy has traditionally focused on conservative applications that use on-device processing to enhance user experiences or introduce extra features, in contrast to more disruptive experimentation seen in companies like Google or OpenAI.
Earlier this year, The Information reported that Apple was exploring LLMs but was cautious about potential unforeseen consequences, such as incorrect information or “hallucinations” (errors) generated by AI chatbots. Recently, Apple also restricted ChatGPT use among its employees. The report also indicated a talent drain of strong AI experts from Apple to companies like Google and Microsoft, which adopt less conservative attitudes towards AI.
Apple will address obstacles and integrate LLM-based features into its offerings
Despite Apple’s efforts to explore AI tools like ChatGPT within the company, significant challenges remain in bringing these tools to Apple products. It remains uncertain how or if Apple will address these obstacles and integrate LLM-based features into its offerings.
Apple’s top AI executive for Machine Learning, John Giannandrea, is currently leading the division in developing chatbot tools similar to ChatGPT. Internally, Apple’s engineers have dubbed this chatbot project as “Apple GPT.” Additionally, Apple’s Senior Vice President of Software Engineering, Craig Federighi, is also involved in the development of these tools. However, both executives seem to have differing views on the direction that Apple should take with AI.
While the specific details of Craig Federighi’s approach to advancing AI were not discussed in the report, there is evidence to suggest that the two executives are not in full agreement on how Apple should proceed. John Giannandrea reportedly wants to witness the extent of improvements that ChatGPT can bring before making decisions on how Apple should integrate this technology. This difference in perspective might explain why Apple has yet to establish a clear strategy on how the speculated chatbot will function for consumers.
Curious to learn more? Explore our article on: Enterprise Wired