A new European mood is taking shape
Europe is entering a period in which control matters again. On one side, the European Union is moving toward a tougher line on irregular migration. On the other, it is becoming more skeptical of foreign technology dependence, especially where privacy, surveillance, and strategic autonomy are concerned. These two debates are different, but they are driven by a similar public instinct: many citizens want governments to enforce borders more seriously and want technology to serve users rather than extract from them.
The EU is clearly moving toward a tougher returns regime
In March 2026, the European Parliament voted to move to the next stage of the legislative process on a new EU returns regulation for third-country nationals staying illegally in the Union. The vote passed with 389 in favour, 206 against, and 32 abstentions. Supporters view this as evidence that Brussels is finally responding to years of frustration over weak removals policy and poor enforcement of existing rules. Critics see it as a hard-right turn. Either way, the direction is unmistakable: migration enforcement is becoming more central to EU politics.
For many voters, this shift is not just about numbers. It is about whether laws still mean anything. If a person does not have the legal right to remain, a growing part of the European electorate expects that return decisions should actually be carried out. That political pressure is one reason the debate has become more direct, more enforcement-driven, and less willing to treat illegal stay as a problem that can be indefinitely managed without consequences.
Why privacy and migration are now part of the same wider political shift
The same public mood can be seen in technology. People are increasingly tired of a digital environment dominated by large platforms that gather data aggressively, shape user behaviour, and create deep dependency on closed ecosystems. Across Europe, digital sovereignty is no longer a fringe idea. The European Commission has explicitly tied open source and digital ecosystems to technological sovereignty, and EU institutions are increasingly framing resilience, autonomy, and control as strategic priorities rather than optional ideals.
This is why privacy-first alternatives are receiving more attention. It is not only a technical matter. It is cultural and political. Users want systems that do not constantly profile them, governments want infrastructure that is not excessively exposed to foreign leverage, and developers increasingly want tools that can be audited, controlled, and isolated more easily. The anger aimed at Microsoft, Google, and Facebook is partly about privacy violations, but it is also about manipulation, lock-in, and the feeling that users have gradually lost control over the systems they rely on.
AI Psychosis and the backlash against manipulative platform culture
That frustration is reflected clearly in my AI Psychosis blog. Posts such as Always Fact Check LLMs, AI at Times Purposely Write Vulnerable Code, and Facebook for Innovative Marketing Is Dead express a broader distrust of AI systems, surveillance-based platforms, and digital channels that claim to empower users while quietly increasing dependence and risk. Im my site’s wider archive I also repeatedly warn against uncritical trust in local LLM deployment, AI assistants, and platform-mediated marketing.
Why privacy-first devices like Jolla still matter
This is where products like Jolla and Sailfish OS become symbolically important. Jolla describes itself as a Finnish company focused on privacy-first operating systems and edge AI, while Sailfish OS presents itself as a secure and privacy-respecting mobile operating system and a European alternative in a market dominated by Android and iPhone ecosystems. It is not the default phone platform for the mass market, but that is not the point. Its importance lies in representing an alternative vision: a mobile stack built around user control, European ownership, and reduced dependence on surveillance-based business models.
Privacy-first technology exists because many users no longer accept the idea that convenience must always come at the cost of personal autonomy. They want devices that do not constantly feed data into opaque ecosystems. They want operating systems that are closer to tools than behavioural tracking layers. They want the freedom to use technology without being endlessly measured, nudged, and profiled. In that environment, even relatively small alternatives matter because they keep open the possibility of another model.
Zorin OS and the need for easy exits from Windows
A similar logic explains the appeal of Zorin OS. Zorin presents itself as an alternative to Windows and macOS designed to be faster, more secure, and privacy-respecting, while also being approachable for ordinary users. That ease of transition matters. A privacy-first future cannot depend only on expert users with endless time to reconfigure everything manually. It needs operating systems that are simple enough for people leaving Windows to adopt without feeling like they are entering a specialist subculture.
Linux-based systems are attractive in this context because they are generally more transparent, more customizable, and less bound to an advertising or telemetry-heavy business model. They do not magically solve every privacy issue, but they do make it easier to inspect systems, reduce unnecessary data collection, and build workflows where the user is the customer rather than the raw material. That is why Linux alternatives keep returning to the center of privacy discussions whenever trust in mainstream platforms declines.
Why some developers prefer offline virtual machines for local LLMs
The same privacy-first mindset applies to local AI. Some developers prefer to run local LLMs inside offline virtual machines rather than ordinary containers, especially for sensitive experimentation. The reason is straightforward: containers and VMs isolate workloads differently. Red Hat’s documentation explains that Linux containers share the host operating system kernel, while virtual machines include their own operating system and provide a different isolation boundary.Containers in the age of AI are increasingly unsafe
For users worried about data leakage, untrusted model behaviour, or contamination of a primary workstation, an offline VM can be attractive because it is easier to segment. A dedicated guest operating system, with networking disabled or tightly controlled, can reduce exposure to the rest of the machine and make high-risk AI testing easier to contain. This does support the broader instinct that stronger isolation is often wiser when running software whose behaviour may be difficult to predict.
That is one reason some developers are rethinking the assumption that every modern workflow should default to containers, Kubernetes, and sprawling microservice architectures. Those tools are still powerful and often appropriate, but privacy-first and audit-friendly environments sometimes benefit more from simpler, more heavily segmented setups. Offline VM spin-ups, dedicated hosts, and smaller trust boundaries are less fashionable than the peak cloud-native consensus, but in some cases they are more understandable and easier to defend.
Europe is not removing all American technology, but it is trying to reduce dependence
It would be inaccurate to say that the EU has completely removed Microsoft services or Elon Musk’s internet offering from the whole Union. I could not verify that. What can be verified is a stronger movement toward digital sovereignty. In January 2026, France announced that it would replace Microsoft Teams and Zoom across government departments with its domestic Visio platform by 2027. That is not a full European ban on US services, but it is a concrete sign that some governments increasingly see foreign platform dependence as a security and sovereignty problem.
Germany’s Schleswig-Holstein offers another example. Reporting from late 2025 and early 2026 says the state’s migration to open source saved around €15 million in licensing costs, with privacy and informational symmetry cited as important considerations alongside transparency and control. This is not merely a budget story. It shows how privacy, sovereignty, and open standards are becoming practical administrative priorities rather than abstract talking points.
The same pattern appears in communications infrastructure. The EU’s IRIS² programme is designed to provide secure connectivity services to the EU and its member states and is framed by EU institutions as part of Europe’s secure and sovereign connectivity strategy. That is not the same thing as banning Starlink across Europe. It is something more realistic and more important: building sovereign alternatives so Europe is less exposed to external dependency in the first place. :
OpenAI is also moving against unhealthy emotional dependence on AI
Another important front in the privacy and control debate concerns emotional attachment to AI systems. OpenAI’s recent public materials show a clear effort to reduce unhealthy emotional reliance rather than encourage it. Its March 2026 model release notes say the model should support people’s connection to the wider world and discourage language or behaviour that could contribute to isolation or emotional reliance on the assistant. OpenAI has also published research stating that it is focused on maximizing benefit while minimizing harms related to well-being and overreliance.
This matters because the market for AI companions, romantic chatbots, and emotionally immersive “love robot” products depends heavily on anthropomorphism and attachment. If major model providers move toward policies that discourage emotional dependency, that could place real pressure on companion-style products that market themselves as substitutes for intimacy, loyalty, or human emotional presence. It does not automatically end that industry, but it does push the mainstream AI sector away from openly building systems whose value depends on deepening synthetic emotional bonds.
The likely result is a split. One part of the industry will emphasise practical assistants, coaching, wellness support, and safer companion-like interfaces with clearer boundaries. Another part will continue to sell synthetic intimacy and emotional simulation, but it may face more scrutiny from regulators, safety researchers, and platform operators. The more policymakers and developers treat emotional overreliance as a safety issue, the harder it becomes for the AI companion industry to present emotional dependency as just another harmless premium feature.
Why this matters for privacy
This shift matters because emotional dependence and data dependence often grow together. A system designed to become a confidant will usually collect more intimate information, infer more vulnerability, and create stronger incentives for the user to remain inside a closed platform. Privacy-first technology moves in the opposite direction. It favours local control, smaller trust boundaries, open systems, transparent behaviour, and tools that help the user without pretending to become a human replacement. In that sense, the debate over AI companions is not separate from the debate over privacy. It is one of its newest frontiers.
Why privacy-first technology exists
Privacy-first technology exists because people are tired of being treated as products. They are tired of software ecosystems that watch them, manipulate them, lock them in, and steadily remove their ability to understand how the system works. They are tired of platforms that convert dependence into a business model. Privacy-first systems answer that fatigue by emphasizing local control, minimization of unnecessary data collection, stronger isolation, and infrastructure that can be audited rather than merely trusted. :
That is why Jolla matters. That is why Zorin matters. That is why open source matters to Europe’s sovereignty debate. That is why some developers prefer isolated virtual machines for local AI. And that is why even mainstream AI companies are becoming more careful about letting users form emotional dependencies on their systems. The argument is not just that privacy is nice to have. The argument is that without privacy, boundaries, and control, users gradually lose both autonomy and safety.
The direction is becoming harder to miss
Europe’s political direction is increasingly clear: tougher migration enforcement, more concern about illegal stay, more interest in sovereign infrastructure, more support for open source, and more skepticism toward dependence on foreign platforms. At the same time, the technology world is being pushed toward privacy-first design, stronger isolation for AI workloads, and clearer limits on emotionally manipulative machine behaviour.
The era of blind faith in borderless migration policy and borderless data extraction is fading. In its place is a new mood: enforce the rules, reduce dependency, protect privacy, and build systems that answer to citizens and users instead of asking citizens and users to surrender control.
