Mental Health Action

The need to develop a state alternative to big tech ‘clouds’^[1] to remove the threat of commercial wealth extraction and control of our public services; and how this could be done. ^[2]   

Martin Blanchard, KONP data working group

Introduction:

As campaigners for the benefits of ‘state funded and provided healthcare able to fully and fairly meet clinical needs’, we are aligned with many other groups of campaigners facing the problems caused by the ever increasing ‘wants’ of extractive private corporations. Now together we face ‘digital capitalism’ as practiced by technology giants using swathes of ‘stolen’ human and science data as a resource to provide infrastructure and develop innovations, to survey and control our behaviours, to help practice war, to enhance their own and their investors’ profits, 

Within the UK  they have moved beyond commercial markets into our essential ‘public’ and governmental services, to find for themselves influence, power and reliable, long term profits .

These are the richest corporations ever seen in human history. They are US ‘big tech’ who have developed  a range of monopolies to maintain their global digital supremacies, and have used their power to help to impose their national politics and their digital ideologies. Some call this ‘digital colonialism’.

We have invited big tech into our state funded public services since 2013 to ‘reduce costs and improve productivity’. That big tech provides, and may make available to us, the infrastructure used for US military hardware could have played a role in this, and it may yet be a stumbling block for any wished for future change in UK policy. However, I present this paper as a briefing, developed from a large body of research by Professor Cecilia Rikap and colleagues, into the economic and technological developments that need to be countered urgently to prevent new data, knowledge and power becoming solely acquired by big tech- with suggestions of how to do this. If we do nothing then, as political economist Cédric Durand points out: 

‘As big tech begin to centralise the algorithmic means of coordination of all public activities’, and ‘as public institutions are rendered incapable of organising society’s needs’- through lack of public expenditure- ‘that task will then fall to big tech, with all the power that brings’.^[3]

PART ONE: OUTLINE OF A PROBLEM AND A SOLUTION

Big tech: their monopolies and power

1. Digital technologies/data/ digital capitalism in innovation have become the bedrock of the US and Chinese economies, their military power and any geopolitical ambitions.

• Digital technologies are disproportionately provided by a small number of corporations Microsoft, Amazon, Google, Apple, Nvidia, Meta, Oracle with other smaller/subordinate companies from the USA, and Tencent, Alibaba, Baidu and Xaiomi in China. In this paper I use the abbreviation MAG where just these first three are the major players.

• Big tech performance in the US economy is currently a major factor in preventing its recession.

• There is an increased realisation globally that where any law stands in the way, then any dominant ‘actor’ in a nation that can protect big tech activities will garner US support. This is even though these activities end up privileging a corporate minority and worsen local and global inequalities and hardships.

• Big tech’s value and power grew over years. Enormous investment in cutting-edge compute for data centres and the leading start-ups^[4], and theft of huge amounts of personal data and knowledge, enabled big tech to develop  ‘chokepoint’ monopoly power^[5] in key segments of the developing digital technology global ‘value-chains’^[6]and  sustained  hugely substantial financial and intellectual rent.

• Perhaps more important is that as well as concentrating among themselves specific segments and resources such as fundamental platforms^[7], data centres, datasets and AI talent to develop AI models, MAG are also able to exert control ‘beyond ownership’^[8]of virtually all the  AI and other digital technologies’ ‘value-chains’ including  healthcare. They steer priorities and extract value from thousands of participating organisations.THESE THREE GIANTS DICTATE EVERYTHING FROM THE CONCEPTION OF NEW TECHNOLOGIES TO THE OPERATION OF DIGITAL MARKETS. 

• MAG dominate the provision of cutting-edge-infrastructure: the space where digital technologies are produced, exchanged and consumed.

• The huge number of participating actors selling and buying on big tech infrastructure DOES NOT, as some may think, make the space competitive, and this is because they must all follow rules about exchange and development solutions dictated by MAG.

• Big tech infrastructure is rapidly spreading its net, becoming the single space for the development and adoption of technologies that are crucial not just for companies of all sizes and the economies, but for  health and social services, and  institutions that make up nation states. All are increasingly compelled to use big tech with the potential to become ‘locked-in’ clients.^[9] This grows big tech power.

1. There is therefore an urgency to produce an alternative to big tech to overcome the severe reduction of (digital) sovereignty and control for both states and people, as has been well recognised  and expressed by the European Parliament, and many in Latin America and India. 

1. If the necessary changes are to occur such that independent states can advance their own digital sovereignty and are able to build just societies in which every citizen has access to necessary services including healthcare, and can live without the need for environmental harm and exploitation, then we must develop an alternative to the ‘value- chains’ controlled by big tech.

1. Unfortunately, current UK proposals to rescue the economy are focussed on achieving  ‘productivity gains’ driven by big tech AI even though the idea of such productivity increases are not verified empirically.^[10] All the while both the social and ecological effects of using AI in this way are invariably overlooked.

1. What is needed is not just more physical infrastructure but to replace MAG infrastructure as the central space for everything to do with digital technologies. This will require the creation of an alternative ‘digital value chain’ or what is also called a (national) ‘stack’^[11] guided by four principles:

1. It is publicly led: as a public and democratically governed digital economy with a ‘public utility’ of public physical infrastructure and public platforms for essential functions such as the development of new solutions, open-source foundational models^[12] and architectures^[13], and ‘public’ marketplaces.
2. It is democratically governed: neither corporate interest nor techno-nationalisms should be allowed to capture and control it.
3. It is people-centred: with prioritised technologies chosen to improve people’s lives and working conditions-this includes through the provision of better public services. A public-led ‘stack’ should confront the current situation of monopolistic wealth- extraction.^[14]
4. It is developed within planetary ecological boundaries.

Even though many feel overwhelmed by the thought of such change, 

IT IS POSSIBLE TO EXPAND DIGITAL SOVEREIGNTY– to obtain ‘the democratic control by states and their people of essential technologies for their lives and self-government. All we need is the willingness, and capability to work together with like-minded others to achieve it’.

PART TWO: ABOUT THE PUBLIC STACK

1.What big tech monopoly power is based on, and the Public Stack solutions:

1. Big tech can capture large numbers of ‘start-up’ companies, as well as ‘the frontier’ of AI.

Transforming what would otherwise be a risky and very high fixed cost into a variable cost is particularly tempting for small organisations. The expansion of any ‘start-up’ ecosystem^[15]  therefore becomes closely tied to those corporations with the necessary services and infrastructure capability, where organisations pay only for the storage, processing power and other computing services they use such as platforms, software and data. Left without any alternative, start-ups can become ‘captured’ as perpetual rent payers, with their entire business models dependent on MAG . Also, at the other end of the industry, because the computing demands required are so huge, larger corporations such as OpenAI, Anthropic and Mistral need to rent cutting-edge MAG infrastructure to develop their frontier AI models^[16] as well. 

Stack Solution:

Governments need to invest more public money in ‘promising’ public start-ups (publicization?). With a public stack able to provide a full-service of tech facilities for the development of those publicly invested ‘start-ups’ and frontier AI models, huge future expenditures of rent on private provision could be avoided and we could maintain control of key technologies. The state may also obtain competitive rental income for the public purse from allowing use by commercial enterprises that follow ‘stack’ rules; indeed quite a few may wish to do this to avoid big tech control.

1. Big tech can take ownership and most of the profit from jointly developed products through use of their ‘asymmetrical marketplaces’ and their ‘panopticon’^[17] view of ecosystems.

Companies using big tech marketplaces end up integrating into a space where the rules are those of big tech, and where their actions are subordinated, and large parts of their profits siphoned off by the giant owner (asymmetry). Also, as the monopoly supplier only MAG  are able to gather technological and intellectual knowledge and use it to control what happens on their infrastructure and across entire innovation ecosystems (panopticon). This knowledge also enables them to ‘cherry pick’ who they help and who they invest in or buy out, so that they can maximise their own profits and they alone can remain at the ‘cutting edge’

Attempted solutions:

AI factories:

A governmental role in initiatives such as the UK’s ‘Nightingale healthcare  AI factory’^[18] will not be able to replace reliance on big tech’s infrastructure. This is because the factories need to use big tech’s infrastructure when working with the huge amounts of data  in longitudinal and/or linked population data analyses, complex genetic analyses, and AI development and training. Again, even giant companies developing AI foundation models with their own processors, such as  Deep Seek and Meta, because of the amount of data involved ultimately depend on big tech’s infrastructure to sell them as a service. 

Decentralised, multiple federated marketplaces:

 As a solution suggested by the proposed ‘EU solution’, the federated marketplaces had difficulty being accepted because  big tech offered a much more convenient service where organisations needed to follow and comply with only a single set of rules. 

Stack Solution:

A single public marketplace in the publicly led stack would replace the undemocratic concentration of decision-making power and oversight with a truly public, open space, for both selling and freely sharing computing services. Again, such facilities could generate income for the public purse from those that wish to avoid big tech control.

1. Big tech capturing public knowledge and using it selectively for maximising profit/rental or to control the entire development of innovation.

Stack Solution:

The need for public knowledge

For any alternative ecosystem or value-chain to flourish and to keep fundamental knowledge ‘open’, ‘foundational’ platforms^[19] must be offered as public services and should remain ‘open source,’^[20]  enabling everyone to build applications and other specific solutions on top of them. This should include the need for platforms that offer essential services to be considered as DIGITAL PUBLIC UTILITIES (note that search engines, e-commerce marketplaces, AI foundation models and social media have already become essential infrastructure of today’s societies).

 There should be a ‘commons’ governed by new public institutions with state, union and civil society representation. 

Further considerations:

A competition layer

The stack should also offer actors of every size and specialisation all the necessary layers for the development of specific solutions in what should be a competition layer. This layer could include open source, collaboratively developed services offered for both free and proprietary applications. Every computing service should be granted the possibility to be part of this layer as long as they are compliant with established regulations and environmental standards,

Such an ecosystem could initially be promoted to provide services for a specific sector, for example healthcare. In that case, the specific competition layer would offer services targeted for that vertical^[21] and the open AI foundation model to be developed should be tailored to that sector. 

Public funding 

This should be used to encourage the open-source community to develop specific services identified as gaps or missing in the publicly led stack. Exploratory, on-demand funding should also be available for concrete proposals on how to improve the stack and make it more sustainable. 

A dedicated Digital Technologies R&D Agency 

This needs to be established to develop and update the necessary architecture of the public stack. R&D on AI and other digital technologies must find a space independent of all production networks controlled by big tech and other corporate giants. Public knowledge networks should focus on an independent science and technology research agenda to consider economic, political, ecological, and ethical priorities and their implications separate from the current one that is driven by the pressures of techno-logical solutionism and market needs.

A Non-Profit International Research Institution

Building frontier models requires hundreds of scientists working together on long-term projects, which is a good reason to establish such an agency dedicated to produce these digital technologies for people and the planet. It must include researchers from the social sciences and humanities in sufficient numbers to ensure that there is adequate consideration of economic, political, social and environmental insights into the impacts of digital technologies. The agency should also have regular dialogue with the publicly-owned organisations managing both physical infrastructure and the public marketplace, and with the final users of the technology – such as healthcare  providers and schools as well as the populations they serve– to collectively identify priorities.

The use of public procurement

It is essential that public procurement of all digital products and services should be channelled through this publicly led stack to assure demand and to guide the development of digital technologies towards social and environmental goals

Whence funding for a PUBLICLY LED STACK

Taxes

These should be drawn in part from: 

• taxing big tech’s appropriation of data and knowledge perhaps by use of digital services taxes or levies on the extraction of free data by large platforms.
• corporations paying fair and adequate income taxes. 

Taking a regional or international approach

The cost of building an alternative to big tech will be substantial, and a regional/ international approach would be the most viable. The shared initiative would also safeguard against attempts by specific governments to dismantle the work being done, especially if the agreement made it totally clear that:

‘Any use of the developed digital technologies for the surveillance, targeting and extermination of their citizens, and those beyond their borders is strictly forbidden’.

Governance, competition and power  

A DEMOCRATIC GOVERNING BODY

The publicly led stack should have a democratic governing body which is publicly funded but autonomous, with a mandate to prioritise communities and planetary good. Such a body must understand that:

• democratic solutions will NOT arrive via unregulated control by a few giant corporations, and nor with unsuccessful attempts to promote competition. For example, the promotion of competition on its own will not prevent natural monopolisation in the several layers of the stack prone to having single firms that are more efficient and ecologically ‘safer’ than two or more providers^[22]– 
• the pursuit of traditional solutions to economic power abuse, such as ‘monopolist break ups’, will at most be an incomplete solution^[23].   
• in search engines and other AI-powered services, algorithms improve through use, and this results in dynamic economies of scale and powerful network effects that seriously reinforce any natural monopoly. 
• many algorithms are endlessly improved with all our prompts, which means thatSOCIETY AT LARGE IS COLLECTIVELY CO-PRODUCING THEM. But at the same time, it becomes nearly impossible to privately challenge the ‘for free’ decades of algorithmic refinement and data accumulation. 
• Key digital services, such as foundation AI models, search engines and social media, must be conceived as, or become public utilities
• Competition policy needs to evolve and take NETWORKS AND ECOSYSTEMS AS RELEVANT UNITS OF ANALYSIS, so that it is possible to consider the effects of ‘chokepoints’ (see Footnote 4) and ‘panopticon’ views (see Problem 3) across a digital system^[24] in the development of monopolies.
• The alternative to competition for those segments of the digital stack that are prone to natural monopolisation or to intellectual monopolisation MUST BE TO REMAIN PUBLIC AND OPEN.

The role of any Publicly led digital stack

A publicly led digital stack will not fully replace big tech eco-systems,  but instead offer an alternative and by doing so, competition layers can be promoted and remain decentralised in the rest of the system (see issues dealing with competition above). In addition, to protect the stack, existing and new regulations must be enforced and enacted such as:

• Big tech and other giants must share the population data that they have extracted. 
• A further public institution could be created as big tech’s steward, and this could operate as a Data Agency that also grants access to publicly held data to those from the public sector working on projects that should remain public and open. All these datasets should be used only for public research and policy on a principle of data solidarity i.e. in which citizens are clearly informed and requested to grant permission for use of their data.
• There must be an obligation to share data in the case of public procurement where every company contracted by the public sector shares with it all the data harvested as part of providing the service, in an editable format.
• The state should be provided with a golden share in the affiliates of large digital corporations operating within it. Such a golden share will expand chances to supervise the deployment of new services, audit algorithms ex ante^[25] and request compliance with local regulations.
• Institutions receiving state funding MUST COMMIT NOT TO ACCEPT BIG TECH FUNDS OR SERVICES for their research and innovation to limit knowledge extraction. 
• The United Nations should be encouraged to foster an assessment and redefinition of digital technology standards aimed at reversing large tech companies’ use of ‘standard setting’^[26] to impose their technologies. 

Empowering people and communities

International cooperation is crucial for expanding digital sovereignty since knowledge expands when it is shared^[27]. Collective solutions are cheaper, more likely to succeed and more environmentally friendly. 

The case of healthcare data illustrates how collective data sharing can generate benefits that are unlikely to emerge from a market- based data economy. With appropriate governance and regulation, centralised healthcare data could play a crucial role in managing the global health care crisis. Cross-referencing healthcare data with socio-economic and environmental datasets, and analysing them through digital technologies, could provide evidence to enhance prevention, diagnosis, treatment and care delivery. These and other collective solutions should be explored at the international level.

Only by expanding digital sovereignty through a publicly led national/ international, ecological and democratic alternative can we:

• remove control from the hands of big tech and place it within democratic, international institutions.
• shape the trajectory of digital technologies so that they serve collective needs and wants of the population while respecting  planetary ecological boundaries.
• address common problems and contribute to reduce inequalities^[28]. 
• create systems that do not impose the same technologies on every citizen but instead empower people to make informed decisions about which technologies they want to use, for what, when and how. 

‘THE MAIN GLOBAL DIGITAL DIVIDE TODAY IS BETWEEN THE GLOBAL MAJORITIES THAT USE TECHNOLOGIES AND THE FEW THAT KNOW HOW TECHNOLOGIES WORK, DECIDE WHICH ONES ARE PRODUCED AND HOW, AND THEN CONCENTRATE THE ASSOCIATED PROFITS’^[29]

ADDENDUM

POLICY RECOMMENDATIONS

• The development of science and technology must be steered to ensure that everyone can use, access, understand and produce digital technologies with full awareness of their environmental, social, economic, ethical and political implications. 
• Proposals such as the digital ID should be developed only and to the extent that they can be hosted at state-owned data centres.
• Better collaboration at the international level: negotiating better deals for essential segments of the digital ‘value-chains’ and collective buying in the case of AI semiconductors could be a way to mitigate partially the rest of the world’s disadvantage (compared to big tech).
• The rest of the world can deepen its integration by developing ‘data pools’^[30] and associated common solutions for applying collective digital solutions. 
• Develop international common standards and regulations for data centres: where they can be installed, minimum quality standards and technologies to be used, taking into consideration the effects on the local population. 

COMMENT  

‘In a world increasingly torn apart by antidemocratic policies and the misuse of technologies, in which the chances of overcoming the economic, ecological and geopolitical crises appear bleak and international cooperation dwindles, there is a unique window of opportunity’. 

‘This is the time for us to act together as international allies to create bold alternatives to the present status quo; to develop publicly led, democratic and people-centred digital stacks that put the rights and needs of people and the planet ahead of private profits, the military and the geopolitical claims of a minority of humanity that undemocratically rules the world. 

The UK together with other concurring countries in Europe, Latin America and indeed many others from both the North and South of the Globe, could decide to develop and share new digital ecosystems that oppose the current giant US and Chinese tech dominated ecosystems. This would be done in hope to bring an end to any race for technological supremacy, a race that chooses to ignore all the social and ecological implications of their digital technologies, and cares little about the slaughter of humanity as it runs its economic and geopolitical course.’  

In such publicly led systems, organisations will be able to participate on an equal basis, sharing and co-creating technologies for a better and safer present and future for all. 

Words of hope from Cecilia Rikap:

‘Creating this alternative is feasible, but it demands political courage to rethink prevailing models of statecraft and to challenge dominant narratives that portray publicly led solutions as inherently undemocratic, insufficiently qualified or stifling of innovation.’ 

So many countries are characterised by high-quality tech research and innovation within their diverse network of public institutions. Building on THESE networks (rather than those involving giant private corporations) with all the willing countries/groupings of countries in the world will be crucial to assure the inclusive sharing of the benefits that can be derived from publicly led stacks. We can have all the capabilities to use digital technology in the pursuit of a more just, more equal, healthier and safer world, and will then also stand at least a chance to defeat not just the threats of diseases and war, but also the threat of incipient ecological breakdown. 

---

^[1] Big tech are major US tech corporations. The ones that are mainly relevant to this paper are Microsoft, Amazon (AWS), Google (MAG) and to a lesser degree Oracle.  ‘Cloud ’is generally used as a euphemism for a data centre with advanced technology able to provide essential infrastructure. Big tech cloud is therefore that which currently only big tech can supply. Throughout this paper I prefer to use the terms ‘big tech’ or ‘MAG infrastructure’.

^[2]  Adapted and developed from: A PROGRESSIVE ROADMAP FOR EXPANDING EUROPEAN DIGITAL SOVEREIGNTY by C.Rikap

^[3] Durand, C. ‘Fragile Leviathan?’ SIDECAR 30^th January 2025 

^[4] young, high-growth, and innovative companies designed to quickly develop and scale a new product or service

^[5] ‘Choke point power’ in digital refers to the ability of a dominant entity—such as a large technology firm, platform, or state—to control, restrict, or monitor critical, narrow nodes in a network, thereby exerting significant influence over the flow of data, goods, services, or financial transactions.

^[6] A ‘value chain’ is the full series of operational activities a company performs—from raw material sourcing to final product delivery and support—to add value, increase efficiency, and create a competitive advantage.

^[7] frameworks—comprising hardware, software, or both—that enables other applications, technologies, or users to build, operate, and interact -such as operating systems

^[8] https://www.common-wealth.org/publications/dynamics-of-corporate-governance-beyond-ownership-in-ai

^[9] where a customer becomes dependent on big tech infrastructure provision for services and data, and finds it too difficult, expensive, or time-consuming to switch to any other provider

^[10] In fact, macroeconomic productivity gains driven by AI adoption are expected to be less than 0.53% over the next decade. Acemoglu, D. (2025) “The simple macroeconomics of AI”. Economic Policy, 121(40): 13–58.

^[11] A ‘stack’ approach to IT infrastructure—encompassing hardware, software, networking, and data storage—designed to ensure an organization or nation maintains exclusive, independent control over its data and digital operations, free from foreign legal, political, or private corporate technical interference.

^[12] large-scale AI models trained on vast, generally unlabelled datasets that are made publicly available for inspection, modification, and redistribution.

^[13] a collaborative, transparent approach to designing systems that are publicly accessible.

^[14]While currently it is rivalries, polarisation and individual gains that are promoted – what is required is digital technologies designed to promote the socialisation of gains and knowledge to expand the common good, with strict regulations on what data can be harvested and used for training AI models.

Streamlined public-sector procurement could serve as an engine for this public-led digital stack with additional financing through measures like levies on the extraction of free data by large platforms. 

^[15] A technology development ecosystem is a dynamic, interconnected network of businesses, startups, investors, universities, and government bodies that collaborate to drive innovation, share resources, and accelerate growth

^[16] A highly advanced, large-scale artificial intelligence system that defines the current state-of-the-art, pushing the boundaries of capabilities in areas like natural language processing, coding, and multimodality.

^[17] Some have called it this due to an operational likeness with Jeremy Bentham’s (1748-1832) ‘panopitcon’ a structure that enabled a few guards to monitor hundreds of prisoners/workhouse labourers and control their behaviours.

^[18] A Uniquely British AI Factory for Healthcare https://www.imperial.ac.uk/news/261693/launching-nightingale-ai-new-foundation-world/

^[19] Large-scale, pre-trained AI systems that serve as the core, underlying base for building specialized, downstream artificial intelligence applications. A term coined by the Stanford Institute for Human-Centered AI in 2021, that refers to models trained on vast, diverse, and often unlabelled data, which allows them to be adapted—via fine-tuning or prompting—to a wide array of specialized tasks without starting from scratch

^[20] software with source code that anyone can inspect, modify, enhance, and redistribute, fostering collaborative development.

^[21] a specialized sector where companies offer services tailored specifically to the unique needs of the particular industry or service niche.

^[22] For instance, several search engines asking the same type of questions would be less efficient than one, as users would need to consult several platforms to find the best answer. 

^[23] The main flaw in the divestment hypothesis lies in its failure to account for the fact that firms operate within (asymmetric) ecosystems and production networks, sharing information selectively. This is not to say that divestments should not occur, but to emphasise that they cannot be seen as the main, let alone only, policy solution to economic power abuses. 

Being  forced to divest assets and break up has currently been discussed for Google in the USA, but it was felt that the new companies would most likely remain under its control. Divested companies can still work together, sharing databases and research results to maximise rent extraction.

^[24] Big Tech and other large tech companies are intellectual monopolies that use their position in those networks to capture data and the knowledge that is produced collectively and turn it into their rent- bearing intangible assets.

^[25] An ex-ante audit is a preventative “before-the-event” evaluation performed prior to the commitment of funds or commencement of a project. It involves reviewing planned actions to ensure compliance, feasibility, and efficiency, often focusing on risk assessment and checking procurement policies before money is spent.

^[26] Standard setting in technology is the voluntary, supposedly collaborative process of establishing common technical specifications, guidelines, or protocols for products and services.

^[27] Knowledge expands when it is shared because it triggers a process of refinement, validation, and collaborative growth, transforming individual insights into collective intelligence. Sharing allows others to test and build upon ideas, uncovering hidden perspectives, reducing duplication of effort, and fostering new, creative solutions that ultimately increase the total knowledge available.

^[28] This should also include those inequalities in internet access as a political priority. However, PURSUING DIGITAL INCLUSION WITHIN THE CURRENT SYSTEM WOULD REPRESENT AN EXPANSION OF DATA SUBJECTS FOR BIG TECH and other large platform companies and the increased power that brings. This is why proposals such as digital IDs, e-payments or any data exchange platform should be developed only and to the extent that they can be hosted at state-owned data centres and should always offer citizens the possibility to op-out.

^[29] Because only this minority retains the whole oversight of the system, they decide the future of the most ubiquitous technologies of our time, ultimately deciding how societies learn and what they learn about. 

^[30] data pools generally refer to centralised, shared repositories where data from multiple sources—such as different departments, organisations, or IoT e.g wearable devices—is gathered, cleaned, and standardised for easy access, analysis, and collaboration.