Elon Musk is wrong — but the risks of AI in your iPhone and Windows PC should 'not be underestimated'

tesla ceo elon musk on the left with a screenshot from chatgpt working on a mac in the next version of the apple operating system right

Elon Musk shared criticism of some of the Artificial Intelligence announcements from WWDC on his social media platform X. The multi-billionaire had previously blasted Microsoft over its AI features too

REUTERS | APPLE PRESS OFFICE | GBN
Aaron Brown

By Aaron Brown


Published: 19/06/2024

- 14:33

Updated: 20/06/2024

- 08:46

"Disable everything that you don’t need to share"

Like it or not, Artificial Intelligence (AI) is coming to every device you own.

When ChatGPT was named the fastest-growing consumer application of all time in early 2023 — reaching 100 million monthly users in two months, a milestone that took TikTok nine months and Instagram over two years to achieve — it kickstarted an arms race between some of the most valuable companies.


Silicon Valley has been exploring these technologies for years. Still, the record-breaking demand demonstrated by ChatGPT pushed companies like Apple, Microsoft, and Google to begin implementing Generative AI and Large Language Models (LLMs) into their products to match the capabilities of OpenAI's clever AI 'bot.

Microsoft launched Copilot — and added a new key on its keyboard dedicated to the AI assistant, the first shake-up to the standard Windows keyboard layout since the Start key was introduced back in 1994. And this week, it released a brand-new category of computer, known as Copilot+ PCs, with an even greater focus on AI features, like the controversial Recall feature that screenshots, reads, and categorises everything you do on your device so that it can be searched like "a photographic memory".

animated gif showing the copilot key leaving the keyboard and twirling in the light

Microsoft has placed such emphasis on its ChatGPT-powered Copilot AI assistant that it made the first change to the standard keyboard layout for all Windows PCs in three decades — a Copilot key

MICROSOFT PRESS OFFICE

Meanwhile, Google Gemini is a direct competitor to ChatGPT that's available across your devices — and could be baked directly into iOS 18 to improve Siri in a forthcoming update from Apple. It's also using Generative AI to overhaul its search engine, writing bespoke answers rather than simply presenting a list of websites.

Meta is bringing its AI chatbot to millions of WhatsApp and Instagram users, with no option to remove the feature that can interact in your Group Chats like an all-knowing virtual friend. Alexa is widely-tipped to get a ChatGPT-like upgrade in the coming months — at a cost.

Earlier this month, Apple announced its long-awaited entry into the fray with Apple Intelligence and the ability to summon ChatGPT using its voice-activated assistant Siri from your iPhone or Mac.

And Elon Musk was not happy about it.

The multi-billionaire suggested he would enforce an outright ban on employees using an iPhone, iPad, or MacBook at his companies over the newly-announced plan to integrate ChatGPT into its next-generation operating systems. The Tesla and SpaceX executive blasted the decision as an "unacceptable security violation" and accused Apple of handing over user data to OpenAI.

Elon Musk was an early investor, writing the AI start-up a cheque for $50 million. However, the 52-year-old has since terminated his relationship with OpenAI and launched his own competitor, known as Grok, modelled on The Hitchhiker’s Guide to the Galaxy that can answer “spicy” questions "rejected by most other AI systems".

Writing on X, Mr Musk said: "If Apple integrates OpenAI at the OS level, then Apple devices will be banned at my companies. That is an unacceptable security violation. And visitors will have to check their Apple devices at the door, where they will be stored in a Faraday cage."

In a further post, he added: "It’s patently absurd that Apple isn’t smart enough to make their own AI, yet is somehow capable of ensuring that OpenAI will protect your security & privacy! Apple has no clue what’s actually going on once they hand your data over to OpenAI. They’re selling you down the river."

But the SpaceX and Tesla executive is wrong about how ChatGPT will work on iPhone. If you can use the new AI features in iOS 18 (it's worth remembering that millions of iPhone owners will miss out) Siri will always check before sending anything to the latest OpenAI model, known as ChatGPT-4o.

There's no "Always Allow" or "Don't Ask Me Again" checkbox built into iOS 18, Apple will doggedly ask whether you're happy to share some limited data with OpenAI every single time. Even if you do decide to use ChatGPT to answer a particular query, Apple has negotiated a sweetheart deal that prioritises privacy.

During its announcement at WWDC, CEO Tim Cook said that users will always need to grant permission before any request is shared with OpenAI. No requests will be stored, IP addresses will be masked, and users on Apple devices will not need to sign up for an account with OpenAI, Apple added.

iphone 15 pro pictured with a prompt from siri asking if youre happy to send something to chatgpt

If you request something that Apple's own Large Language Model and Generative AI models aren't equipped to handle, Siri will request throwing the request to the latest version of ChatGPT. But it will always ask permission before sending any data to the OpenAI system from your device

APPLE PRESS OFFICE

Dr Ilia Kolochenko, who serves as an Adjunct Professor of Cybersecurity at Capitol Technology University, has described the warnings from Elon Musk as "exaggerated".

Speaking to GB News, Dr Kolochenko said: "Elon Musk’s warning seems to be exaggerated and unsubstantiated — unless he has access to some internal information that is not yet public.

"Having said this, whenever an internet-connected object, be it an iPhone or Tesla automobile, shares something with its vendor or its partners, privacy and security risks should not be underestimated.

"Importantly, information sharing risks are not necessarily related to the now-booming use of AI, for instance, many mobile devices or laptops with pre-installed software share technical and other data with their vendors by default under different pretexts. While such practices are unlawful in most cases in the US, UK and EU, some vendors still try to find 'creative' bypasses to legal rules and stealthily collect invaluable metrics and more sensitive data from their unwitting customers."

Device owners should always be aware of any personal data shared from their devices with third-party firms — but there's no innate threat with these incoming Artificial Intelligence models.

When pressed about what device owners can do to elevate these privacy concerns, Dr Kolochenko told us: "To avoid sharing your personal data, preferences, internet activities and geo-location with third parties, make sure that you understand what, when and how your device and all installed applications — like Siri or ChatGPT — send to the Internet.

"Disable everything that you don’t need to share, this shall reduce 90% of most common privacy risks."

Dr Kolochenko is far from alone in thinking that Elon Musk has misunderstood the implementation of ChatGPT coming to iPhone, Mac, and iPad in the coming months. Annette Zimmermann, a VP Analyst for US research and consultancy firm Gartner, described Apple's approach as "a holistic AI strategy".

"Apple revealed in my view a robust two-pronged AI strategy," Ms Zimmermann said. "On the one side, you have the personalised and contextualised experience based on Apple’s proprietary GenAI technologies going across first-party apps and third-party apps and powering new capabilities in Siri.

"And on the other side, you can draw from — as Tim Cook put it – 'world knowledge' with ChatGPT. But still keeping privacy in mind by building a mechanism into the ChatGPT integration that asks for consent every time I start an interaction with ChatGPT."

This viewpoint chimes with that of Jack Hayes, a leading thought leader on AI, who told GB News: "We are moving into a new era of technology and to dismiss AI as inherently unsafe for use by the public would be misleading."

"Its use on Apple phones and devices should not automatically be seen as controversial, chiefly because the company is not nearly as dependent on advertising revenue as its rivals," Mr Hayes adds. "It makes a virtue of the ways it protects the security of its users and Apple would be making a monumental error if it were to undermine its own business model by misusing data gathered by AI in some way."

However, as mentioned by Dr Kolochenko, there are risks to all of these internet-connected features. Recall was unveiled as a flagship new feature for Copilot+ PCs to encourage huge numbers of PC owners to upgrade to a brand-new laptop, like the Surface Laptop 7 or Galaxy Book 4 Edge. But the AI-powered functionality has caused a storm of controversy for Microsoft.

An ex-Microsoft engineer branded the feature as a "disaster" when he revealed that transcriptions of everything on-screen, including bank details, sensitive text messages, and more stored in a plain text document that could be viewed by anyone with access to your PC.

“The issue with Microsoft’s proposed use of its Recall technology is that, with AI taking regular snapshots and analysis of your activities, is the information gathered going to be of great use to users, or will this information find its way to advertisers and commercial partners?," Jack Hayes tells GB News.

"Some might see it as allowing your own device to spy on you, which highlights the need to ensure that data is held and used securely. Users need to care about the information they share when using AI and ensure they are in control of their choices, while businesses need to ensure that the rights of individuals are protected."

The UK data protection watchdog has confirmed that it's "making inquiries" with Microsoft, Apple executives ruthlessly mocked the security flaws in the feature, and billionaire SpaceX CEO Elon Musk urged millions of followers on X to kill the feature. Faced with this immense backlash, Microsoft has delayed Recall for all Copilot+ PC owners and promised to bring extra protections.

It's a relief to see those warnings from privacy experts have forced Microsoft, worth an estimated $3.3 trillion, to backtrack and fix the flaws. By all accounts, Elon Musk seems to be wrong about the implementation of ChatGPT, but he was correct about the dangers of Windows 11's Recall functionality.

AI isn't innately good or bad for privacy, but each new update should be assessed on its own merits and how it treats your personal information. With so many Artificial Intelligence features on the way — we all need to be better equipped to protect ourselves.

You may like

{% if context.post.roar_specific_data and context.post.roar_specific_data.affiliate_post %} {% elif %} {% endfor %}