Apple disables ChatGPT to prevent confidentiality leaks! A large-scale version of Siri will be upgraded and launched soon

This article is reprinted with permission from AI new media Quantum Bit (public account ID: QbitAI). Please contact the source for reprinting.

ChatGPT has just been launched on iOS and has become a global sensation, quickly dominating the App Store.

But then the news came out that Apple’s own employees were not allowed to use ChatGPT , and it had been several months.

This matter immediately sparked huge discussion.

You know, since the ChatGPT trend came, Apple has almost remained "invisible". After all the effort, the iPhone has embraced ChatGPT, but now it seems to have poured cold water on it.

Some people have already started to mock:

Is Apple afraid that its employees will find AI actually useful?

So they're going to use their own? Hopefully not like Siri.

Eh? Who knows?

With this big news, news that Apple is researching its own large models has also spread, and Siri will soon launch new AI features .

Cook’s latest public statement: Generative AI technology has great potential, but there are problems that need to be solved , which also indirectly confirms their progress in AI.

So it’s not that Apple is not planning on releasing it, maybe it’s just that the release time has not come yet.

Concerns about data leakage

According to the Wall Street Journal, Apple prohibits employees from using ChatGPT and Copilot.

A well-known Apple product tracking reporter also further revealed on Twitter that ChatGPT has been on Apple employees’ banned list for several months.

Such a decision is not uncommon.

On the one hand, ChatGPT itself has been controversial in terms of data usage .

Initially, all user chat records except API users were used to train and improve model performance, which meant that user data was at risk of being leaked.

In fact, such an accident did happen in March this year.

A bug caused some users' conversation titles to be leaked, and others could see this information on their own interfaces. At that time, this bug temporarily closed ChatGPT.

Last month, OpenAI officially launched new regulations to further protect user data security:

Users can prevent their personal data from being used for training by officials by turning off the chat history function.

On the other hand, for security reasons, many large companies have introduced policies requiring employees not to use ChatGPT.

For example, Amazon asked employees in January not to enter code into ChatGPT after it was reported that they had discovered that ChatGPT would imitate Amazon's internal code.

JPMorgan Chase restricted the use of ChatGPT by employees worldwide in February due to compliance concerns.

Even more bizarre is that Samsung once said that there were three data leaks due to ChatGPT within 20 days.

It is reasonable for Apple, which has always been known for its emphasis on security, to take similar measures.

After all, if an employee accidentally enters information about a confidential project into the system, that information could be leaked to OpenAI.

Others said that the output results of ChatGPT may still be erroneous, which may lead to bugs, data loopholes, etc., so it is reasonable for large companies to ban its use.

However, since ChatGPT became popular a few months ago, its improvement in work efficiency and creativity is undeniable.

Many people have incorporated it into their workflow, including big names like Terence Tao.

New York public middle schools, which initially banned students from using ChatGPT, also lifted the ban within two days.

Therefore, some people are inevitably worried that Apple’s closure of ChatGPT will stifle its own innovation capabilities.

It is worth mentioning that Amazon, which once banned its employees from using ChatGPT, has recently joined the big model trend and released its own big model platform.

This also makes people begin to speculate that the big companies are not unwilling to embrace ChatGPT, but just want to embrace their own ChatGPT?

In fact, there is real news about Apple’s latest language model.

Siri is reported to launch new AI features soon

According to 9to5Mac, Siri will soon launch new AI features.

The new technology is code-named Bobcat, and it was revealed as early as March this year that it was already being tested.

First tested on the latest tvOS 16.4 beta, the new framework is called Siri Natural Language Generation , and then this feature is introduced to Apple's other operating systems.

But the revelation shows that it has only realized telling jokes with Siri on Apple TV. In addition, it is also testing how to use language generation for timers.

At the same time, the leader of this new project was also revealed.

John Giannandrea is currently Apple's senior vice president of machine learning and AI strategy, reporting directly to Cook.

He joined Apple as early as 2018, overseeing the company's entire AI and ML strategy as well as the development of Core ML and Siri technologies. Before joining Apple, he spent eight years at Google, leading the machine intelligence, research, and search teams.

Cook's statement in a recent investor conference call also indirectly confirmed that they are doing related research and development, but the time for release has not yet come.

Previously, Apple employees had complained about Siri's low efficiency. The reason why Siri gradually declined was due to internal team chaos, slow decision-making, and cumbersome code , which seriously hindered the development of Siri and AI.

The more fundamental reason lies in technological research and development and judgment.

For example, the core functions and technologies currently demonstrated by ChatGPT were once rejected by Apple executives.

Like the ability to have long conversations, they think it can make the conversations get out of control and be "fancy".

To prevent Siri from speaking nonsense, Apple chose to have a human team write the answers in advance, and also repeatedly refused to allow users to provide feedback on Siri's answers to questions, resulting in the development team being unable to understand the limitations of the model.

Although it cannot be confirmed that Apple’s ban on ChatGPT is related to its own development of AI tools.

However, many people are probably looking forward to the launch of new AI functions of Apple Siri.