Last month, France decided to cease Google Meet, Teams, and Zoom from the administration by 2027. This echoed plans by a German state, Schleswig-Holstein, to switch their system software in 2024 to an open-source alternative such as Linux, Nextcloud, or LibreOffice from the Big Tech proprietary software with years’ preparation. These plans were recently regarded as successful. Denmark also followed Germany’s steps in 2025, aiming to use LibreOffice as an alternative to Microsoft.

While these European governments’ policies arose out of concerns for cost-saving, national security, and state sovereignty in the face of the US’s increasing aggressiveness, there are important underlying values at stake for all of us. Dangers are not only to digital privacy, but also to our very human agency, which enables us to act as individuals and as societies. 

I will first illustrate how the values and agencies mentioned above have become so vulnerable under the global attachment to Big Tech digital service, then I will briefly introduce open-source digital tools, and explain how they can empower us to restore our agency and privacy. 

Digital Privacy 

Rodion Kutsaiev

One of the most broadly discussed issues related to Big Tech’s proprietary software and services are the threats they pose to data privacy. This was made worse after the Large Language Model (LLM) went viral. For instance, Meta (formerly Facebook) recorded 20 data breaches between 2016 and 2019. One of them was the Cambridge Analytica scandal, in which a political consulting firm had collected identifiable personal data from around 87 million Facebook profiles through Facebook’s Open Graph platform, and constructed corresponding psychological profiles. Meta was fined $5 billion by the Federal Trade Commission in 2019 and £500,000 by the UK Information Commission, but despite the attempts of governments to impose laws and regulations, their impact on pushing Meta to conduct effective changes remains insignificant. This is clear through the recurrent data breaches and Meta’s responsive attitude towards the incidents.

If we focus on the user’s side, a much more complex phenomenon can be identified, namely the privacy paradox. That is, although people may claim privacy does matter for them, their actions often indicate the contrary is true, where they end up in a “trade off” sacrificing their personal information for (better) services. Such a mentality is probably not unfamiliar to most of us, which I will argue later is a pseudo dilemma. Given the public’s original preference on privacy is not able to effectively hold them back from providing their personal data, it is not surprising that Meta’s data breaches constantly happen; their monthly active users grew rapidly from 2 billion in 2017 to 3 billion in 2022.

The problem here is stark. While our reality is increasingly intertwined with the digital sphere, the effects of privacy loss and data exploitation online filters itself back into our everyday life. Putting aside its high-level conflicts with the right to privacy that are supposed to be guaranteed by international law, our loss of control over our personal information renders us extremely vulnerable to technological service providers, data buyers, or thieves. We are curtailed from the knowledge about how exactly our data is used and secured, asked to trust providers, regardless of the endless data leak scandals and illegal data collections. Even upon requests for your own personal data, there is little chance you would see everything, nor the most crucial part, which is how your data has been used. Like the data protection activist Paul-Olivier Dehaye, who in 2016 made and later successfully proceeded a complaint about Facebook to the Irish Data Protection Commissioner. He received a response from Facebook 15 months after the initial request, when they refused to provide the analysis process on his data due to reasons such as “huge technical challenges”.

In facing such systematic exploitation, current regulations, legal actions, and privacy advocacy remain treating the symptoms but fail to address the roots. Our resistance is either insignificant, unable to create a change in behaviour, or remains post-hoc, by seeking accountability in the law, while harm has already been done and which remains unable to prevent recurrences. Such harms range from identity theft crime to construction and consolidation of mass surveillance systems. The latter, I argue, is a great harm to us all, in authoritarian and democratic countries, conducted by governments and key non-governmental actors – the tech firms.

Human Agency 

As Shoshana Zuboff accurately identified in The Age of Surveillance Capitalism: the Fight for the Future at the New Frontier of Power, by collecting and extracting users data ranging from sensitive information to usage data such as time length of staying in a post, the platform services providers can (1) repackage the data as products selling to advertisers to conduct predictive target advertising; and (2) optimising their learning algorithm to predict users preferences, to enable advertisement on their platform becoming more accurate, so to enhance users attachment on their platform, namely “personalisation”. While the will of us, as the concrete living being, did not matter in the commercial process but our behavioural patterns are, data which even users did not intend to share are often used to proceed predictive analysis, trapping us in serving others to take advantage on ourselves by reproducing our demands (or invoking our consuming desires) with predictive advertisements. 

Again, the above might sound like common knowledge to us nowadays. What is considered one of the most important insights from Zuboff is the problematic ubiquitous quality of these mechanisms; that they seek to confine us in an enormous, infinite time cage, and cease our “right to the future tense”. In other words, our present digital reality is shaped by the replication of our past preferences and opinions. This only acts to consolidate our past choices and communities. We are subjugated by the goal to engender total certainty, jeopardising our possibility to act otherwise, or to put it simply, to make changes. 

The example of predictive policing brings this out of the abstract and into a daily, worrying reality. The concept of predictive policing here includes but is not limited to “the data-driven, profiling, and risk prediction and assessment systems used by public entities”, as defined in an Amnesty International UK report.

Under Surveillance

Background relationships between governmental actors and Tech firms are an important context to these recent debates. 

As Zuboff illustrated, after the 9/11 attacks and US national security concerns reached their peak, the US and European governments started to seek cooperation with tech companies to develop mass surveillance technology. In 2003, for example, Google started “customizing its search engine under special contract” with the Central Intelligence Agency (CIA) to oversee “top-secret and sensitive but unclassified intranet for CIA and other IC agencies”. A Wikipedia-like intelligence intranet, known as the intellipedia, was thus developed, supported by the Google system, to allow sharing information among agents from different organisations. Later on in 2007, the intelligence search engine ICREACH was recorded in use for “collect[ing], process[ing] and stor[ing] vast amounts of communications metadata related to worldwide intelligence targets”, purported to predict future “threatening” behaviours. 

Recently, along with the widespread fanaticism on AI, the use of predictive surveillance technology in policing and cooperation between Silicon Valley organisations and governments has become increasingly common. Since 2021, The UK Metropolitan Police Service (MPS) has been using risk terrain modelling (RTM) licensed by Simsi to target geographical criminal hotspots and so to effectively allocate their policing resources. In 2025, the UK police force confirmed a contract with the US surveillance tech company Palantir to build all-encompassing profiles of individuals, including one’s philosophical beliefs, health records, biometric and genetic data, political opinions and race etc., to predict future crime. Unsurprisingly, the metropolitan police has also been using Palantir’s technology to monitor and “identify potential shortcomings in professional standards” within themselves. It must be noted that the costs of these “efficiency-enhancing” measures are insurmountable. 

Setting aside extreme totalitarian cases – like the Chinese government conducting brutal predictive governance with the surveillance technology supplied by US Tech firms, including IBM and Microsoft – the reproduction of cognitive social categories embedded in these algorithmic technologies are not only reinforcing existing biases and further marginalising minority groups. When combined with their black-box characteristics, they are also disabling critical thinking, limiting our ability to take part in the analytic process, and feeding us uncritical “answers”, by predicting what we asked for. In the context of predictive policing, ungrounded arrests or detentions, which in the past law enforcers would have been responsible for, are now blended into the non-transparent mechanical algorithmic process. This obscures the accountability of law enforcement. We may not know who should be blamed in a particular case, as law enforcement decisions will be systematically reduced to the gears of mechanical predictive policing. All agency will be dissolved. 

Marvin Meyer

Along with the soaring popular reliance on AI algorithm technology, these technologies are maturing to provide more and more “accurate” predictions for “our future”. We are sliding towards a stifling totality of patterns, with fewer and fewer options for alternative futures. Buying what the advertiser predicts you would buy, but which you did not have in mind in the first place; dump-scrolling for hours due to the successful constant stimulation by a “prefect” algorithm; or self-censoring your content to accommodate an unreasonable algorithm logic, that you believe to be so, in order to maximise your exposure. 

But there are ways out. Here we introduce the following pragmatic solutions: to phase out Big-Tech owned proprietary software with open-source software, to recover our agency and privacy. 

What is open source software? 

An open source software is a “software that features freely available source code, which users may view, modify, adopt, and share for both commercial and non-commercial purposes.” The development of open-source software should aim to be transparent, collaborative, and decentralised. It is collaborative because it allows the public “collectively reviewing, testing, documenting, and patching code”; it is transparent because the coding source is open to the public and whenever new security issues regarding the code are raised, everyone can contribute to identify and fix it swiftly; and it is decentralised because no one own the code and potentially everyone can use it for development. The code mostly appears with a licence that illustrates the conditions that you are allowed to use the code, for instance, the MIT license just requires the derived version of the code must contain “the original copyright notice and a copy of the license itself”.

The existence of open-source software contains several important messages: 

  • Privacy and convenience are not necessarily a trade-off-relationship. You can have both.
  • Software that requires you to pay is not necessarily more secure than those which do not. In the case of proprietary and open-source software, it can be the other way round. 
  • There are a lot of communities supporting the transparent public digital sphere, and resisting Big Tech domination. 

While it is noteworthy that although the code developments are decentralised, an open-source service as an end-product might be still centralised. However, they still share the transparency characteristics which are crucial. 

The following lists some open-source software and tools for your reference. (These are decentralised services, with the option to host local servers, if not explicitly noted.)

Social Media: Bluesky (centralised service with algorithm but is able to bridge to Fediverse), The Fediverse Network (including Mastodon, Pixelfed, Peertube, WordPress, Lemmy) 

Office Suite: Libre Office (centralised service), Nextcloud (with cloud service as well) 

Video Conference: Jitsi Meet (centralised service) 

Message App: SimpleX, Signal (centralised service) 

Operation System (OS): Linux 

Email: Disroot, Proton (centralised service) 

For more alternatives, visit AlternativeTo and search for services with an “open-source” tag. 

Reluctance to switch software is understandable. One reason may be that you have already uploaded so much personal information to Big Tech services that switching or not seems that it would make no difference. However, we affirm that preventing your future information from being absorbed into the algorithm does make a difference; you are joining the fight to lessen the predictive power of black-box algorithm. 

Another reason might be the opportunity cost of changing everything to open-source, which can be quite high in some situations. Therefore, we call for a phasing out period, instead of switching entirely. In reality, Schleswig-Holstein’s successful reformation took years of preparation before being implemented. Getting familiar with open-source tools gradually is crucial to sustainable replacement. 

Activists Without Borders recommends the following: 

  • Take a small simple step by looking into an open-source alternative to one of your used apps 
  • Share your privacy or agency concerns with your friends 
  • Spread this article 

Together, we can resist the asymmetric predictive power placed on us and restore our human agency.