ADVERTISEMENT
Go back to
Home » My Kolkata » Lifestyle » Apple is reportedly working on a privacy-first AI chatbot

artificial intelligence (AI)

Apple is reportedly working on a privacy-first AI chatbot

The development could be on track for a significant AI-related announcement next year

Mathures Paul | Published 21.07.23, 05:41 AM
Apple BKC in Mumbai, Apple's first offline retail store in India

Apple BKC in Mumbai, Apple's first offline retail store in India

File picture

Apple has been incorporating AI in its apps for years now, like in the Photos app to recognise subjects and through Siri but the trend, at the moment, is generative AI, like ChatGPT, DALL-E, Bard, Microsoft’s Copilot features in Office 365. So why not create its own generative AI? According to a new report from Bloomberg, the company is developing artificial intelligence tools to take on OpenAI, Google and others. The report says that the world’s most valuable company has created a chatbot that some engineers internally refer to as “Apple GPT”. Apple hasn’t made any comments on the development.

The development could be on track for a significant AI-related announcement next year. Apple appears to have built its own framework for training generative AI tools, known as Ajax. Like with any tool from the company, the focus will be on improving the privacy of the new technology.

ADVERTISEMENT

The internal rollout of the chatbot continues but it requires special approval for employees to access. Bloomberg reports that any output from the chatbot can’t be used to develop features meant for customers. The chatbot is being used to help with product prototyping.

To give its AI department a boost, Apple hired John Giannandrea in 2018. Giannandrea, who is now Apple’s senior vice-president of machine learning and AI strategy (reporting to CEO Tim Cook), previously led AI and search at Google. The AI project at Apple is being led by him and Apple’s senior vice-president of software engineering, Craig Federighi.

In recent months, Tim Cook has spoken briefly about how the company is “looking at closely” at generative AI tools. At the same time, Cook expressed concerns about AI products during an earnings call in May, saying there are “a number of issues that need to be sorted”.

The news comes close on the heels of Meta’s announcement that its LLM, LLaMA 2, will be accessible on Microsoft’s Azure platform. Microsoft, of course, runs OpenAI’s GPT model on its Bing search product.

Meta CEO Mark Zuckerberg said that he planned to provide the code behind the company’s latest AI technology to developers and software enthusiasts around the world free of charge. “When software is open, more people can scrutinise it to identify and fix potential issues,” said Zuckerberg in a post on his Facebook page. The latest version of Meta’s AI has been created with 40 per cent more data than the previous version.

If Apple has indeed built its own framework for training generative AI, it could be used to make the company’s existing products and services better, like having a conversation with the HomePod the way one does with ChatGPT and Siri may offer answers from the web in a more conversational way.

Mathures Paul

Last updated on 21.07.23, 05:41 AM
Share:
ADVERTISEMENT

More from My Kolkata