Microsoft’s AI Obsession Is Harming Customers (Here’s Why)

Copilot PC

Microsoft is riding the profits wave with AI, but I think consumers will suffer as a result of this obsession. Here’s why.

Until ChatGPT, I thought AI products consumers could use with any level of real functionality were years away. I was wrong. Microsoft clearly knew better than I did. It invested a significant amount of money and resources into becoming a market leader in AI through its OpenAI partnership. Now, Microsoft seems almost obsessed with AI, and I’m not sure that’s a good thing.

Copilot is the result. Integrating AI into Windows and Office products makes sense for Microsoft. Even so, in my opinion, the speed of integration is happening too quickly and with not enough oversight. The end result is us (the consumer) being harmed.

Microsoft’s energy usage doesn’t match up to its climate ambitions

Technology companies like Microsoft are some of the largest energy consumers in the world. It owns huge data centers, which need a lot of electricity and water to run. It needs more data centers to handle the demand. The construction of those data centers is costly to the environment.

An example of a data centre

The end result is an increase in global emissions of 29.1% from 2020 to 2023. Microsoft’s claim to be sustainable is admirable, but it doesn’t match up to the facts. The demand for its cloud services, including its AI products, will continue to grow as it integrates Copilot into more products and services.

In the future, AI might be powered through 100% clean energy and construction. In the here and now, however, Microsoft needs to do more to show that the AI gold rush isn’t something that will damage the environment the same way Bitcoin is.

Slapping AI certification on Microsoft Windows PCs is harming consumers

If the gold rush began with ChatGPT’s release in November 2022, that means that we’re only 19 months into it, as of this writing. If that’s the case, then how can Microsoft already have AI-focused PCs that are obsolete?

PC manufacturers have been claiming the benefits of PCs and laptops with NPUs for several months now. These AI-focused devices were supposed to be future-proofed against future AI developments, allowing consumers to continue to take advantage of future developments.

Microsoft announced the new Copilot+ AI PC range

Except, with Microsoft’s new Copilot+ certification, those ‘early’ AI PCs aren’t powerful enough, with NPUs that range between 10-20 trillion operations per second (TOPS). Only newer devices with this certification, with NPUs that deliver at least 40 TOPS, will be allowed to use new AI features in Windows and third-party products.

You could blame this on the device manufacturers alone, but as Chris Hoffman at PC World points out, Microsoft has been pushing these devices itself since the start of 2024. It controls what devices are classed as certified for Windows, after all.

Early adoption is always risky, but no PC should be obsolete within months. As my colleague Brian Burgess suggests, it’s a good idea to hold off until the technology improves. Time to count down the days before the Copilot+ certification is superseded or updated.

Copilot Recall is the biggest security risk imaginable

If I had concerns before about Microsoft’s AI focus, they were mild before I heard about the biggest one of them all: Copilot Recall. This feature takes a screenshot of everything you see and do on your PC. It’s a database of everything you’ve ever done, that you can search through and, well, recall.

An example of Copilot Recall being demonstrated on the Microsoft website

That means taking a screenshot every time you type in a password. It means taking a screenshot every time you sign in to your online banking. It means taking a screenshot when you visit any website, open any app, or perform any task. Any file you’ve edited or deleted, any secure messaging app you’ve used, any private email you’ve sent—visible.

No malware could be as powerful as something that Microsoft bakes into its own operating system. Microsoft claims that it is secure, but researchers have already proven that Copilot Recall data can be accessed by Admin-level users in plain text.

As far as I’m aware at the time of publication, there’s no Recall-targeted malware on the market at the moment, but you can guarantee that it’s coming. There are also serious privacy issues to consider, especially in sensitive business environments, but equally at home.

This is the last feature that Microsoft should have ever created. No matter the safeguards, in my opinion, it is a risk too great.

Microsoft’s OpenAI partnership makes money but risks everything

Microsoft owes the lead it has in the AI market to its partnership with OpenAI. It has invested billions of dollars into the (on paper) non-profit enterprise, and as a result, it has integrated versions of OpenAI’s GPT models into Bing Chat initially, and then Copilot.

This partnership allowed Microsoft to integrate OpenAI’s AI model into Bing in February 2023, only a few months after ChatGPT was first released. This allowed it to ride the wave of publicity that came with it and set itself up as a market leader in AI. Improvements from OpenAI have flown into Microsoft products since. 

ChatGPT and Copilot on a smartphone, alongside Bard from Google

OpenAI’s reputation is mixed, to say the least. It has suffered from several public battles, from employees being forced to sign non-disparagement agreements that prevent them from criticizing the company, to ugly battles about the risk that AI might develop sentience and harm humanity.

It was this risk that forced the OpenAI board to sack Sam Altman, OpenAI’s CEO. That led to Microsoft announcing that Altman would join Microsoft to lead the AI team there, only for Altman to be back as OpenAI CEO within days. Instability isn’t great in any company, but certainly in one as important to Microsoft (and the AI community more widely) as OpenAI.

The partnership with OpenAI is clearly profitable for Microsoft. It has given it a commanding lead in the AI space compared to its competitors. However, as the internal battles at OpenAI show, there are huge risks to its reputation. If OpenAI continues to have very public bruising battles, Microsoft’s reputation will suffer as a result.

Microsoft should think carefully

I’m not a luddite. I understand the improvements that AI can make to our lives. In my opinion, however, Microsoft has prioritized “getting ahead” in AI above all else and, as a result, consumers will suffer. 

It won’t take long before the first generation of certified, Copilot+ PCs become out-of-date. Microsoft has integrated AI features like Copilot Recall without giving it serious thought or security testing. The partnership that the company has with OpenAI, a billion-dollar business inside a research organization, is risky to its reputation. The environmental demands from AI services on the environment will increase.

I don’t suggest that Microsoft should do nothing and let AI pass it by as it did the mobile smartphone market. However, as a market leader, Microsoft needs to do something it won’t want to do: slow down and think carefully. AI won’t wait, but if it lets ambition take over, it might find itself driving off the edge of a cliff with OpenAI, facing down regulators and angry consumers alike.

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *


To Top