Ceding control: How Copilot+ and PCs could make enterprises beholden to Microsoft

The personal computer revolutionized society by putting the power of technology directly in the hands of individuals. The openness and modularity of PCs allowed users to customize their machines, create what they wanted without restrictions and maintain control over their digital experience.​​ However, this open spirit that once defined the PC era has been gradually eroding, giving way to a new age of closed, tightly controlled computing.

The decline of open computing

In recent years, we’ve seen a steady decline of the openness that defined the PC era. It started with the shift to less upgradable laptops and mobile devices, led by Apple’s incredibly successful iPhone and MacBook lines. While these products boasted sleek designs and tight hardware-software integration, they came at the cost of user upgradeability and repair rights.

Apple’s walled-garden approach extended to software as well. The iOS App Store model, which Google soon emulated with Android, gave Apple unprecedented control over what software could run on its devices. This “curated computing” paradigm, sold as a security benefit, began acclimating users to a world where they no longer had full sovereignty over their machines.

The rise of cloud computing further eroded user control. Google was at the vanguard here, shifting core productivity apps like Gmail, Google Docs and Google Photos entirely into the browser. This made robust offline functionality a thing of the past and our personal data ever more accessible to cloud providers. Chrome OS took this philosophy to its logical endpoint, producing machines that were little more than conduits to Google services.

It was against this backdrop that the integration of cloud-based AI assistants like Siri and Google Assistant began to fundamentally alter the relationship between users and their devices. Now, instead of a tool that solely responds to your commands, your device proactively shapes and influences your interactions. Helping you became indistinguishable from guiding you in directions that aligned with the assistant maker’s priorities.

But it wasn’t just the PC that was undergoing a transformation from open to closed. The internet itself, once a radically decentralized network where any node could communicate with any other on equal terms, was becoming increasingly centralized under the control of a few giant companies. Google, with its 95% share of the search market, was already the gatekeeper of the internet’s information long before the rise of AI. 

With the integration of its Gemini AI directly into search results, Google now has an even more profound intermediary role, constantly interpreting and filtering our access to online knowledge. Similarly, Facebook’s deployment of Llama AI as an ever-present “assistant” in its social networking apps means users are now subjected to a constant stream of AI-driven suggestions and interpretations, eroding individual agency in our online interactions.

Microsoft ushers in 1984 with Copilot+ PCs 

Microsoft, seeing the writing on the wall, began to aggressively remake Windows in this mold. With Windows 10 and 11, what was once the bastion of open computing became an OS that was increasingly opaque, update-driven and cloud-centric. The advent of AI-powered PCs threatens to be the final nail in the coffin of the open paradigm exemplified by the traditional PC.

Microsoft’s recently announced “Copilot+ PCs” represent the company’s most aggressive push yet towards an AI-driven, cloud-dependent computing model. These machines feature dedicated AI processors, or “NPUs” (neural processing units), capable of over 40 trillion operations per second. This hardware, Microsoft claims, will enable “the fastest, most intelligent Windows PC ever built.” But there’s a catch: the advanced capabilities of these NPUs are tightly tethered to Microsoft’s cloud ecosystem.

Features like “Recall,” which continuously monitors your activity to allow you to quickly retrieve any piece of information you’ve seen on your PC, and “Cocreator,” which uses the NPU to aid with creative tasks like image editing and generation, are deeply integrated with Microsoft’s servers. Even the new “Copilot” key on the keyboard, which summons the AI assistant, requires an active internet connection. In effect, these PCs are designed from the ground up to funnel users into Microsoft’s walled garden, where the company can monitor, influence and ultimately control the user experience to an unprecedented degree.

This split-brain model, with core functionality divided between local hardware and remote servers, means you never truly own your PC. Purchasing one of these AI-driven machines equals irrevocable subjugation to Microsoft’s digital fiefdom. The competition, user choice and ability to opt out that defined the PC era are disappearing before our eyes.

Big tech’s anticompetitive utopia 

Microsoft’s current trajectory is particularly alarming in light of the company’s past antitrust troubles. In the late 1990s, Microsoft faced a major lawsuit from the U.S. Department of Justice and 20 state attorneys general, who accused the company of abusing its monopoly in the PC market to stifle competition, particularly in web browsers. 

At the heart of the case was Microsoft’s bundling of its Internet Explorer browser with Windows, which prosecutors argued gave Microsoft an unfair advantage and hindered consumers’ ability to use alternative browsers. After a lengthy trial, a federal judge ruled that Microsoft had indeed violated antitrust laws. Though an appeals court later overturned some of the initial judgments, the case resulted in a settlement that imposed restrictions on Microsoft’s business practices and required the company to share its APIs with third-party companies.

The case was a defining moment in the history of the tech industry, establishing that even the most powerful companies could be held accountable for anti-competitive behavior. But in the decades since, Microsoft has only grown more dominant, and the parallels between its past conduct and its current strategy with AI-powered PCs are striking.

But the core of its empire emerged unscathed, and in the decades since, Microsoft has only grown more powerful. With the company now leading the charge into an AI-driven, cloud-centric computing model, the parallels to its past conduct are troubling. 

The government probably won’t save us this time

Unfortunately, the chances of a robust regulatory response seem slim this time. Big Tech’s lobbying power has reached new heights, with the industry’s biggest players spending more than $28 million on lobbying in the first quarter of 2024 alone. This represents a nearly $5 million increase compared to the same period in 2023, underscoring the industry’s growing influence in Washington.

This surge in lobbying spending is particularly striking when viewed in the context of the past few years. In 2023, the largest tech companies spent around $68 million on lobbying, a slight dip from the $70 million spent in 2022 but still a significant increase from the $65 million spent in 2021. Meta and Amazon led the pack in 2023, spending $19.3 million and $17.8 million respectively. Google, Apple, and Microsoft each spent over $9 million.

Much of this lobbying is focused on shaping the regulatory landscape around AI, as companies jockey for positions in this increasingly lucrative field. OpenAI, for example, registered to lobby for the first time in 2023 and spent $260,000 on issues like cloud computing, cybersecurity, privacy and AI-related legislation. But the lobbying efforts also serve to cement the dominance of these tech giants more broadly, making it harder for upstart competitors to gain a foothold and for regulators to rein in anti-competitive practices.

The business peril of AI PCs

For businesses, the rise of AI-powered PCs like Microsoft’s Copilot+ line represents a double-edged sword. On one hand, the promise of enhanced productivity and creative capabilities is alluring. However, the price of these benefits is a level of vendor lock-in that should give any IT department pause. By tightly coupling advanced functionality to proprietary cloud services, these machines make it incredibly difficult for businesses to switch providers or maintain control over their own computing infrastructure. The more a company comes to rely on Copilot+ features, the more beholden they become to Microsoft’s ecosystem and the harder it becomes to extricate themselves without major disruption and cost.

It’s true that businesses and consumers can still use other AI models and applications, either online or by installing them locally. However, these alternatives will not be as deeply integrated into the operating system as Copilot is on Microsoft’s platform. This means that using a different AI will involve additional hurdles and a less seamless experience compared to the built-in Copilot features.

Much like the antitrust concerns raised over Microsoft’s bundling of Internet Explorer with Windows in the 1990s, the tight integration of Copilot into the OS gives Microsoft an inherent advantage over competing AI providers. Users who want the convenience and efficiency of an AI assistant that’s woven into the fabric of their computing experience will be naturally steered toward Microsoft’s offering.

This dynamic is already playing out in the consumer space. While Mac users can download and use the standalone ChatGPT desktop app, Windows users who want an equivalent experience are pushed towards using Copilot, as there is no standalone ChatGPT app available for Windows yet. This disparity further consolidates Microsoft’s position and makes it harder for users to opt for alternative AI models without sacrificing convenience and functionality.

As businesses and individuals become increasingly reliant on AI assistants for their daily computing needs, the power to control the default, most deeply integrated option becomes immensely valuable. By making Copilot the path of least resistance on Windows, Microsoft is leveraging its operating system dominance to shape the future of how we interact with computers – and to ensure that future happens on its own terms.

But the risks go beyond mere lock-in. The always-on, always-watching nature of features like Recall also represents a profound security and privacy threat. Sensitive business information, from confidential documents to private employee communications, could be continuously uploaded to Microsoft’s servers without clear visibility into how that data is being used or who has access to it. The potential for leaks, breaches, or misuse is immense.

The AI models powering these features are black boxes, their decision-making opaque and unaccountable. Businesses may find their own operations being shaped by algorithms they don’t understand and can’t control, optimized for metrics that align with Microsoft’s interests rather than their own.

For businesses operating in regulated industries or handling sensitive personal information, the compliance implications alone are staggering. How can a company ensure it’s meeting its legal obligations around data protection when it has effectively ceded control of its computing environment to a third party?

The allure of the Copilot+ line and similar offerings from other tech giants will undoubtedly be strong. But businesses need to think long and hard about the Faustian bargain they’re making by embracing these machines. In the short term, productivity may rise. But in the long run, they may find they’ve sacrificed their autonomy, security and flexibility for a shiny new set of digital shackles.

The wiser path for businesses is to invest in open, interoperable solutions that they can control and customize to their own needs. This may require more upfront effort, but it’s an investment in retaining sovereignty over the lifeblood of any modern organization: its data and its computing infrastructure. As the old saying goes, if you’re not paying for the product, you are the product. With AI PCs, even if you are paying, you may still be the product – and that’s a risk no business can afford to take lightly.

The future of computing freedom

So what can we, as users who value technological self-determination, do in the face of this shift? On an individual level, the most direct response is to vote with our wallets. We can reject devices and services that compromise our autonomy, and support open hardware and software alternatives wherever possible.

But realistically, individual action will only take us so far in the face of the tech titans’ consolidated might. To truly preserve and extend the open computing model, we need coordinated efforts to build and promote viable alternatives to the walled gardens of Apple, Google and Microsoft.

This is where organizations like FUTO (Fund for Universal Technology Openness) have a critical role to play. By dedicating resources to support the development of open technologies and user-respecting software, FUTO and its ilk represent our best hope for a future where we retain agency over our digital lives.

One of FUTO’s key strategies is to fund open-source versions of important technical building blocks like AI accelerators, ensuring they remain accessible to a wide range of actors. They’re also working to make decentralized software as user-friendly and feature-rich as the offerings of the tech giants, to reduce the appeal of convenience-for-control tradeoffs.

None of this will be easy. The lure of shiny objects like Copilot+ PCs is strong, and the might of the tech giants is vast. But the stakes are too high to concede defeat. At risk is nothing less than the foundational promise of personal computing: that these phenomenal machines would empower us as individuals, not yoke us to the whims of a few mega-corporations.

The era of the truly open, user-sovereign PC may be fading, but it’s not gone yet. By making mindful choices about the technology we use and support, and by rallying around the work of organizations like FUTO, we can keep the flame of digital freedom alive. The battle will be long, but it’s one we can’t afford to sit out. The soul of computing itself hangs in the balance.

If you value your digital freedom, hold on to your old PCs. When loaded with open-source operating systems like Linux, they may be the last refuges where you can still use technology on your own terms, without the constant surveillance and control of the tech giants. These machines, combined with decentralized, user-respecting software, represent our best hope for preserving a corner of the digital world where individual autonomy still reigns supreme.