IPO’s consultation on balancing between the rights of AI industry and human creatives: what is the problem to be solved? Part 1

by

 

Between December 2024 and February 2025, the UK’s Intellectual Property Office (IPO) ran a public consultation on how to achieve a balance between the rights of human copyright holders, whose works, often taken from the internet, feed the training of Artificial Intelligence (AI)  models, and the right of the AI developers to develop their industry.

The IPO set out the problem by describing the issues faced by the two sides as:

  1. the problem for the copyright holders is that they are ‘are finding it difficult to control the use of their works in training AI models and seek to be remunerated for its use’;
  2. the problem for the AI developers is that they are ‘finding it difficult to navigate copyright law in the UK’.

Hence, the IPO sought views on ensuring that a set of twin Intellectual Property (IP) objectives is met. The first of these is providing sufficient ‘control’ of copyright holders on their material so that they may be fairly compensated for the use of their IP when used for training AI; the second is ensuring that the ‘access’ of AI developers to large, often copyrighted, datasets is not hampered so that ‘investment’ and ‘innovation’ in the AI industry are also protected. The IPO acknowledged that the debate on copyright and AI ‘has concentrated on large, commercial, generative AI models’ (IPO 2024), which are most likely to affect the creative industries.

However, such a framing of this complex and contentious issue, though commonsensical,  stands in danger of overlooking inequities developing towards another group of stakeholders – those users of the internet who provide, what the IPO has termed, ‘low-value data’ (IPO 2024). The economic interests of this group do not seem a significant part of this IPO consultation as they generally are not in most IP-centred debates on the issue. Yet these Users provide a significant amount of the revenue flowing to AI developers from their daily activities on the internet and on other digital media (Image 1).

Image 1: Algorithms surveille Users even as they provide free services. From collecting data to improve User experience, they move to utilising this data for influencing behaviour so that the AI owners may sell advertisement space by matching advertisers with Users and assuring success for advertisers.
Image © 2025 by Sasmita Sinha. All rights reserved

Changing uses of IP in the AI landscape

The reasons may have to do with the traditional concerns of IP protection as they have evolved over the past 350 years. The development of intellectual property protection in its modern form is mostly traced back to England’s Statute of Anne of 1710. The statute recognised the need to protect the rights of creators of intellectual products, such as books, to profit from their originality, skill and labour (Deazley 2008). The books were deemed to be the property of their authors due to being ‘the Product of their Learning and Labour’ (The statute of Anne, 1710, as cited in Deazley 2008, italics added).

This view of intellectual property is important because until the rise of certain kinds of AI, this ‘internal’ quality in the creation of one’s intellectual property was precisely what gave an individual the right to profit from it. Intellectual property emanates from within the mental processes of its creator(s) and becomes recognisable as private property in law once it is inscribed in a tangible medium, such as on paper or a canvas.

Currently, however, there are AI whose development and functioning is more akin to traditional industrial processes as shown in Image 2, which describes both the creation and functioning of an algorithm at Facebook. Both these processes utilise intellectual and behavioural input from the Users of Facebook in order to create Intellectual Property owned by the company, bringing it substantial growth and profits.

Image 2: One of Facebook’s algorithm that develops from collecting and analysing User data in order to then
direct feeds to the individual that are beneficial for the company’s growth though not always beneficial for the User
Image © 2025 by Sasmita Sinha is licensed under CC BY 4.0

In such cases, it would seem that AI development uses individual User’s data more like a traditional industrial activity. For example, house-building uses raw materials like clay that are baked into the bricks, which are ultimately transformed into a final product that does not retain any immediately identifiable trace of the clay in its original form. Despite their lack of visibility, however, such raw materials have to be paid for in a way that the Users’ data feeding the Facebook algorithm does not.

Using a post-Marxist approach, Morreale et al (2024) have discussed such daily activities by Internet Users as the creation of surplus value, which is finally harvested by platforms like Google and Facebook. According to the authors, the Users perform tasks like those shown in Image 1 above in the form of  ‘micro-tasks … [acting like] … unwitting labourers: individuals who are unaware‘ that their seemingly innocuous activities like creating playlists and rating books are then harnessed and utilised for AI training by these companies. The authors call these practices ‘labour exploitation’ because, they argue, that the ‘technology companies unilaterally extract surplus value from individuals to train AI’.

From ‘productivity’ to distributive justice

These ruminations on the nature of what constitutes intellectual property worthy of protection and financial rewards become significant in the context of AI discussions like the IPO’s consultation since, in the past decade, governments around the world have intervened actively to facilitate the AI industry’s access to large datasets for free or at relatively low cost. The primary driver of such interventions has been an almost non-negotiable emphasis on productivity gains.

However, in the context of the IPO consultation, it is notable that the consultation text cites the Global AI Index 2024, which shows the UK in the 4th position globally and as a world-leading power in AI development and innovation. As Hilty et al. (2021) elaborate, governments are expected to use policy initiatives to drive forward innovation in sectors where there is a clear identified need for innovation to solve an identified social problem. Going by the data cited by IPO, there is no such need in the UK at the moment in the AI development industry. Writing more generally, Hilty et al (2021) also agree that ‘on a general level one can presently observe that AI innovation appears to be thriving’ (p. 62).

Instead, the volume of concern is rising at the helpless exposure of individuals to, what essentially amounts to, a mining of the human individual for corporate profit. The increasing power- and economic inequalities ushered in by AI deployment are a cause for concern too. Placed in the context of the phenomenal per capita economic growth that has taken place globally in the past 50 years or so (Data Page: GDP per capita 2023), in spite of a significant jump in the world’s population at the same time (which has more than doubled to its current 8 billion people in the same period [PM 2025]), the emphasising of productivity and competitiveness seems misplaced.

It is, therefore, suggested in this series of blogs that reframing the issue as one primarily involving distributive justice within society rather than one centred on promoting productivity captures the current problems to be solved in a much more meaningful manner. Of course, IP law has traditionally been more concerned with rewarding innovation than matters of justice. However, by exploring the definition of ‘intellectual’ in IP laws, I propose that a widening of this definition, particularly via the lens of distributive justice, may be a much more effective approach to a just balancing of the rights of all the stakeholders in the current AI-development scenario, than staying within established ideological parameters, which may erase the contributions of the most vulnerable group and deny them their valid economic share of the AI pie.

 

References

Deazley, R. (2008) ‘Commentary on the Statute of Anne 1710', in Primary Sources on Copyright (1450-1900), eds 
L. Bently & M. Kretschmer, www.copyrighthistory.org [online resource].

Intellectual Property Office (IPO), 'Copyright and AI: Consultation', December 2024, https://www.gov.uk/government/consultations/copyright-and-artificial-intelligence/copyright-and-artificial-intelligence#c-our-proposed-approach
[online resource] (accessed 20.9.2025). Contains public sector information licensed under the Open Government Licence v3.0.

Morreale, F. et al. (2024) 'The unwitting labourer: extracting humanness in AI training', AI & Soc, 39, 2389–2399,  
https://doi.org/10.1007/s00146-023-01692-3 [online resource].

Population Matters (PM), https://populationmatters.org/the-facts-numbers/ [online resource] (accessed 20.9.2025).

'Data Page: GDP per capita', part of the following publication: Max Roser, Bertha Rohenkohl, Pablo Arriagada, 
Joe Hasell, Hannah Ritchie, and Esteban Ortiz-Ospina (2023), 'Economic Growth'. Data adapted from Bolt and 
van Zanden. Retrieved from
 https://archive.ourworldindata.org/20250915-110645/grapher/gdp-per-capita-maddison-project-database.html 
[online resource] (archived on September 15, 2025).

 

Does public perception of the innovator as a human individual promote corporate oligopolies?

by

Does public perception of the innovator as a human individual promote corporate oligopolies?

 

Innovation is a term that has been increasingly popular and even ‘fashionable’ in social sciences since the 1970s (Mohr 1976, p.1). With the establishment of the World Intellectual Property Organization (WIPO), and especially since the globalization associated with the rise of digital economy, its use and importance have entered the popular parlance too. Mohr (1976) points out that innovation as a term is ‘associated with improvement’ and the ‘act of innovation’ is highly ‘laden with positive value’ (1976, p. 1).

This positive association is hugely important for driving policies in a certain direction, where benefitting the innovator is considered an indisputable public good. Significantly, then, the image of the ‘innovator’ is of utmost importance in garnering public support behind these policies. A simple collocation check for the word ‘innovator’ yields 40 sentences, of which the majority, shown in the table below in the column on the right, suggest a meaning that conjures up a human individual rather than a corporate entity. Yet in policy and academic literature and debates on the subject, the meaning more readily conjured up is that of a commercial entity. Edler et al. (2015) write in a review of academic literature on the link between IPRs and innovation policy that ‘ Intellectual property rights are major means for firms to appropriate the value of their inventions’ (p. 6, emphasis added).

Table showing collocation for the word ‘innovator’–suggesting corporate v individual as its meaning

(Click on the link above to download and view the table)

Throughout their discussion of the link between IPRs and innovation policy, the meaning of the innovator as corporate entities predominates. As IPRs are a crucial instrument for allocating benefits of innovation, this meaning becomes utterly crucial in navigating the tension that the authors identify between ‘the monopoly function and the diffusion function’ (p. ) that the IPR regimes contain. While the public support for allowing the benefits of innovation to go to innovators reflects an understanding of rewarding creative or problem-solving individuals or small enterprises, it often fails to fully appreciate the huge policy tilt being created in favour of large corporate bodies.

Economically perhaps the most significant, but also the most controversial of the IPR instruments is the patent. Edler et al. (2015) summarise that in policy research, design and evaluation IPRs are taken as a barometer of the innovative capacities of companies, regions and countries even though ‘patents as indicators of innovation are highly contested’ (p. 6). This is because patent registration often reflects strategic considerations rather than only ‘the protection of essential invention (Blind et al 2005)’ (Edler et al. 2015, p. 6). The question, therefore, arises as to whether the exclusive protections offered to companies by patents and the benefits accruing from them, converting into a substantial increase of their economic power, are really rewarding a human capacity that the original IPR system set out to protect and which the public view of an innovator applauds protection for.

Similarly, the word innovation’s ‘positive value’ that Mohr (1976, p. 1) pointed to derives from its underlying association with ideas of ‘improvement’. What is significant to emphasise is that in the public imagination the meanings of ‘improvement’ still relate to a betterment of the fundamental human condition. Innovations, across board, are taken to make better those essential needs that apply to all humans and hence cannot be denied on the basis of power, politics or other contentious issues. Innovation, the implication is, binds us together across divisions and its value cannot be questioned. For example, the article at the WIPO site here https://www.wipo.int/en/web/ipday/2017/ten_innovations     discusses 10 innovations that better human lives by making them ‘healthier, safer, and more comfortable’. Health, safety and comfort are fundamental human needs and anything that is said to improve them carries a visceral appeal.

Some of the ideas listed in the WIPO article fill one with admiration like the one about the “auto-exploration bra”, created by a young man, motivated  by his mother’s late cancer diagnosis, to provide a better equipment that women can use on a regular basis to improve the chances of breast cancer detection early. In fact, the medical technology field has been a burgeoning field of research with ‘Medical-related patents [seeing]… a 76.3% increase in granted patents, from 30,429 in 2023 to 53,648 in 2024‘ but it is difficult to determine how many of these are by individuals when the field is dominated by big companies like Medtronic, Johnson & Johnson, Cilag AG, Becton, Dickinson and Company, and Boston Scientific. Under the circumstances, innovations like the “auto-exploration bra” get headlines and create a warm emotional envelope of positive public sentiment that benefits all medical changes listed as innovation. All these innovations, however, are not only improvements in a certain medical field but are also, if not primarily, economic instruments that shift economic and other power equations between the patient/customer and the manufacturer/professional involved in prescribing their use. The respect for innovation and scientific research can often obscure the societal, economic and environmental impacts of new products and evade deeper scrutiny balancing the magnitude of betterment promised with that of the shift in control of human and natural resources that organised corporate innovation puts in the hands of the large companies.

Writing in 2020, in the context of digitisation, Veale pointed out that ‘considerations of whether ‘computing’ was useful in a given context’ were often not given adequate space in deliberations on increasing use of technologies. Instead, simply ‘increasing the availability or intensity of digital use’ was considered as ‘improvement’. Hence there is a process of collectivization at work here whereby all ‘innovators’ are generally perceived as individuals and the rewards they received through IPRs and other policies are considered fair compensation for a valuable human capacity contributing to a desirable meeting of basic human need. For large scale adoption of a product or service, its innovative value itself should not be reason enough. Some of the other questions that might be asked are:

  1. What is the process in place right now for serving this need?
  2. What will be the percentage of improvement? Will it be the same in all contexts? If not, are there thresholds that may be devised below which the adoption of the product may not be of much value?
  3. What will be the environmental cost of producing this article?
  4. How does the improvement in numbers compare with the environmental cost?
  5. Will data from large numbers of people be collected? If so, who will be in the best position to create further marketable ‘innovations’ from it at the fastest speed? Will it be the same patent owner? If, yes, how will it affect the power balance in that field of enterprise? Will it create monopolies or oligopolies?

The vital importance of this last question cannot be stressed enough because this is where the collectivization of meaning of ‘innovator’ interacts materially with power distribution in society. In capitalist societies, which would currently encompass almost the whole world, the discourse of innovation validates swelling inequalities. By triggering human sympathies reserved for remarkable human achievement, the discourse of innovation supports a rolling back of the State in order to incentivize innovative individuals to solve human problems and be rewarded economically for it without cumbersome bureaucratic processes.

However, the result has often been that it has proved increasingly difficult to challenge the power shift to big business that this discourse facilitates. Complex factors in modern capitalism dictate the need for constant innovation of new products by businesses to keep making profits (Figure 1 below). Since business profits provide jobs, welfare, infrastructure and defence money through taxation, national governments are increasingly hobbled in their role of providing societal checks on the power of big business. The larger the government spend, the more the governments must rely on businesses, in effect, creating a power nexus that leaves little freedom for any interests that may diverge from those desired by parties to this nexus.

Figure 1: Modern capitalism and the power of organised innovation

The resulting decimation of effective State power, when coupled with this discourse, places inordinate control over vital natural and societal resources in the hands of larger private multinational corporations, which, in reality, dominate patent grants. Innovation is no longer only, or even primarily, rewarding individuals. Moreover, the logic of an urban industrial society denies most individuals any direct production relationship with nature because natural resources are viewed primarily as ‘standing resources’ (Heidegger 1953) to feed industrial needs.

Inevitably, this logic places governments and ordinary citizens in the power of commercial organisations working for their own private profit rather than for public welfare, but still becoming intrinsically linked to it. Industrialised innovation in the hands of organisations, which include some of the richer universities, that can devote entire departments and large chunks of corporate budgets simply to pursuit of innovating for the market disadvantages individual innovators, who find it hard to compete with this mighty machine.

Edler et al (2015) point out that government measures are often aimed to lead directly to ‘subsequent innovation’ (p. 7), further reducing the spontaneous element in innovation activities and putting large organised innovation at an advantage that grows cumulatively with each successful innovation. Both within and between nations power relations become concentrated on feeding increasing amounts of natural and other resources to organisations that can churn out products faster and faster. In a majority of cases these happen to be large multinationals, relying on increasing automation that further adds to their capacity to innovate. The snowballing of economic rewards for these companies that occurs from this process leaves few challengers to their power in an international governance space that increasingly resembles a Hobbesian state of nature.

Particularly in the case of developing countries, the importance of Foreign Direct Investment (FDI) being promoted by both theorists and practitioners, means that national governments have to agree to skewed agreements to attract FDI. Under Bilateral Investment Treaties (BIT) mostly Western investors are privileged because the BITs, in a majority of cases, do not include issues of importance for the local populace ‘such as labour, migration, environment and human rights clauses‘. Further, the State–Investor dispute mechanisms operate away from the public gaze and scrutiny in ways that most often allow ‘foreign investors [to]… elude the justice systems of host countries and challenge host states before arbitral tribunals‘.

Changing public perception of the terms innovator and innovation, therefore, to reveal their close links to issues of power at both national and international levels is of crucial importance. The public ought to be able to choose between forcing a change of direction between the two options of reverting the reality back to the original meaning of benefitting more individual-centred innovation or erecting power centres in society that effectively check and balance the ill effects of the currently over-weaning powers of the corporates in the name of supporting innovation. The IPR regime can be amended to facilitate these changes too.

References

 

ANAQUA (2025) ‘Anaqua analysis of USPTO patenting statistics 2024,’ Patent Management, 30 January 2025, available online (accessed 20.7.2025).

Downs, Jr., G. W. and Mohr, L. B. (1976) ‘Conceptual issues in the study of innovation,’ Administrative Science Quarterly, vol. 21, no. 4, Dec., pp. 700–715.

Edler, J., Cameron, H. and Hajhashem, M. (2015) The intersection of intellectual property rights and innovation policy making – a literature review. World Intellectual Property Organisation: Department for Transition and Developed Countries of the World Intellectual Property Organization, World Intellectual Property Organization, Geneva.

Marchini, J., Morales, J. and Roffinelli, G., ‘Conflicts between Latin American countries and transnational corporations : The challenges of the region in the face of asymmetrical investment treaties,’ Investment Treaty News, 30 July 2018, IISD, available online (accessed 20.7.2025).

Veale, M. and Borgesius, F. Z. (2021) ‘Demystifying the Draft EU Artificial Intelligence Act,’  SocArXiv. July 6. doi:10.9785/cri-2021-220402.

WIPO (2017) ’10 innovations that are improving lives,’ World Intellectual Property Organization website, available online (accessed 20.7.2025).

 

Image courtesy Photo by Jason Goodman on Unsplash

 

Governing technology: Can the subaltern speak on social media platforms?

by

Governing technology: Can the subaltern speak on social media platforms?

Recently the question of technological governance, in particular the governance of artificial intelligence (AI), has become a central issue in a certain kind of political circle. This is the circle that believes in using politics for the promotion of general welfare. Worryingly, in the fragmented and contentious political discourse now most often conducted through social media platforms, its voice does not always carry through to the people.

In September 2024, the United Nations adopted the Global Digital Compact, which was adopted by 193 member states as part of the UN’s Pact for the Future. One of the aims of the Global Digital Compact is to

This aim seems to have been in discussions, though mostly as a subtext, in this week’s contentious pronouncements regarding freedom of speech on social media platforms and the internet more generally.

This blog post argues that this aim of the Global Digital Compact needs to be made a regulatory priority because, coupled with the uses of AI by social media platforms, the recent pronouncements have the potential to harm freedom of speech while presenting certain individuals, parties and platforms as its defenders.

Free speech and the role of AI on social media platforms

The role of social media in fuelling divisive debates had already been an issue of concern, but the use of AI by social media platforms has heightened concerns from responsible actors across board that the capacity of AI to actively manipulate user behaviour can undermine free speech rather than provide a free space for its exercise.

AI is now deeply integrated into nearly all parts of the work and leisure lives of most people, at least in the global North, and increasingly in the global South too. However, its hidden nature and ‘behind-the-scenes’ uses are not always apparent to those interacting with it.

This is the context in which one must view the recent trend amongst certain right-leaning groups to justify, in the name of freedom of speech, types of speech and arguments which would normally be frowned upon in measured political debates. For example, in his recent comments, the Reform UK leader Nigel Farage has focussed on what he termed as the return of free speech (‘”free speech was back” on X’ ) on social media, especially X (formerly Twitter).

A similar concern for removing ‘bias’ was shown by Meta, the owner of Facebook, who has also decided to do away with ‘factcheckers’ for Facebook, Instagram and Threads (used by 3 billion people worldwide) apparently to remove ‘bias’ and a practice that had become a ‘tool to censor’. Both the owner of X, Elon Musk, and the owner of Meta, Mark Zuckerberg, have recently moved key businesses to Texas from California, the latter considered a more ‘liberal’ state in the United States.

How can AI manipulate user behaviour, reaction or speech?

AI is a computer algorithm or programme that can read patterns if presented with very large amounts of data. Its special features are its speed and scale at which it can handle data, far surpassing the capacities of even the most able human individuals in this regard.

On large internet platforms, including social media platforms, AI models are routinely deployed. They are often programmed to give the owner business certain types of information about the users of the platform. For example, some can work out what makes the users angry (and now many of our devices even have the capacity to read our faces) and put that information together with ‘data’ about our ‘likes’ and ‘dislikes’ we indicate on the platform, the amount of time we linger on a movie ad, and a plethora of other such micro-data over a period of time to work out our personalities and preferences.

The AI not only collects user-generated data but also analyses it constantly. With its speed, and often using the psychological and other expertise that it has been programmed with, it can ‘read’ the minds and behaviours of users who feel they are acting freely in a safe space when acting on the internet. This information has been used successfully to control and manipulate the behaviour and emotions of users once the AI has worked out what makes the user happy, or upset, or angry. They are then shown more of the content which elicits big strong reactions.

These key players’ aim is not to be social with us and engage in conversation, as your neighbour or friend might do if they wanted to find out about you, but to exert control in a way that benefits the people and businesses who have the money, knowledge and resources to create and deploy bigger and more powerful AI algorithms all the time. This control mostly takes the form of ‘nudges’, using the user’s preferences to show them advertisements, news feeds or even friends’ posts that are likely to put them in a certain kind of mood to take a certain type of action, which is ultimately beneficial to the AI owner. This may be a ‘nudge’ or creating digital environment that makes you buy a car or a book or can potentially make you write a divisive and angry political post that you may not normally have written.

This control is not exerted simply by knowing how to develop an AI model. With all the expertise in the world, it would be very hard for a dreaming idealist to sit in their living room or access the local library to create an AI model to match the capacities of Google’s Kubernetes model with its over a trillion ‘parameters’, because that would require millions of dollars to do. In the current AI race, the big players are far ahead and difficult for start-ups, whom they routinely absorb to grow even bigger themselves, to catch up with.

The economic and the political in AI deployment

The justification for the pervasive AI analysis, or ‘surveillance’ of the users, is often given as improving the customer experience by making it easier for the customers to find information that they like or find useful. Importantly, the big technology companies, American and Chinese in the main, have been supported by their governments in developing these technologies in order to create more ‘growth’ or wealth.

However, the United States Senate’s hearings in the elections of 2016 about the Cambridge Analytica scandal heard evidence that suggested powerfully that both in the American elections and in the Brexit campaign on 2016 such technology was used to manipulate the voting behaviour of users.

The cries of ‘disinformation’ and manipulation have been raised from other quarters too implicating Russian, Iranian or other expertise in controlling the behaviour of users through AI, which works at a speed that is overwhelmingly fast compared to human cognition and understanding.

Under these circumstances, does freedom of speech exist in a meaningful way on social media platforms?

Resources and power in the digital space

Increasingly, we are being forced to go digital in all walks of life, including work, health, play, communication, studies or politics. More and more free apps and facilities are being made available that make actions easier and more ‘pleasurable’. A lot of them, however, come from a small group of corporate sources. Even smaller companies creating free content use ‘free’ software and platforms made available by the tech giants such as Google workspace or Amazon’s cloud facilities. Amazon, Google and Microsoft together control almost 60% of the cloud infrastructure.

Despite their ‘virtual’ nature, digital technologies run on hardware. The ‘cloud’ is not a free-floating soft, woolly amalgam of gases and light but a huge ‘black box’, or rather banks upon banks of them, controlled and guarded zealously by the big corporates that control them. Entering a facility housing databanks is probably equivalent to trying to enter Fort Knox in the lore of old.

There is good reason for this. Databanks control the arteries of global interactions in almost everything we do digitally, which is not only the internet use on mobiles or laptops but also Internet of Things (IoT), wearable, smart devices and home gadgets. Almost all users, from governments to ordinary individuals, depend on the space and expertise made available by these corporations. All the data generated through all such interactions, whether about satellites in space or my weekly groceries shop done online, or your school’s lessons made available through the internet, your chat with your mother in another country or the businesses hosted in the Amazon Marketplace, is stored here.

Accessing and utilising Big Data

‘Big Data’ or data on a topic in huge quantities, for example, 10 years of my WhatsApp chats, when used to train AI models, provide invaluable insights and patterns of thought and behaviour that can be used in many different ways, leading to more product ‘innovation’ on a scale for Big Tech that is difficult for newcomers to match.

Of course, data is not always free to use as it is the Intellectual Property (IP) of its owner. However, in many cases the user has little real choice in handing over their data to the app/platform they are interacting with.

Further, uses of ‘secondary data‘ are allowed for ‘research’ purposes, specially for scientific research, without permission by most countries’ IP laws. It has often been argued that very large tech companies have used their knowledge and financial resources to hoover up data,  use for their AI-training purposes, even in ways not allowed by the law, and then destroyed it. AI training needs such vast amounts of data that it is difficult for owners to know or to prove, particularly as the AI output often does not directly reflect the training input. Big Tech has also sometimes used the strategy of starting with ‘research’ purposes to develop their AI and then switching to commercial, proprietary uses since they have the deep pockets to fight lawsuits should such needs arise.

Power of agenda-setting

With such control that is invasive, often affecting humans at ‘subliminal’ (below the level that we are conscious of) level, and with little power in the hands of humans to resist, either because they have few or no other options to the digital, or because their personalities are being read with a view to manipulating them, exercising freedom of speech in any meaningful way can become problematic.

Moving large parts of human social, political and economic interactions to the digital space, without a transparent, accountable and binding global regulatory architecture in place first puts tremendous agenda-setting powers in the hands of platform owners. This power undermines the aim of the Global Digital Compact to ensure independent, fact-based and timely information for all.

In the context of the moves by Musk and Zuckerberg, it is clear that these parties have their own agendas, which are not purely economic. To take just one example, Elon Musk cut almost 80% of jobs on taking over Twitter, and promoted a culture of ‘long hours at high intensity, enforcing a demanding work environment‘, which was his view of what is worthwhile in a human being’s contribution to a work environment. The debate in philosophy on what is the purpose of technology is old and long: is it there to make work more humane and human-centred, so that the ordinary person may earn a living without breaking body and soul, or is it there simply to make more and more money for an abstract notion of ‘welfare’ where an increase in wealth of a country overall, even if it is only in the hands of the top 1%, is counted as progress or welfare?

To promote his agenda of a gruelling work ethic, Musk, a billionaire, absorbed reported losses in revenues of the platform X, formerly Twitter, and has increasingly politicised his approach to debate, now taking it to government-level in the most powerful country in the world. With an ability to control the users behind the scenes, and a definite agenda to promote, the questions to ask are:

  1. how far is Musk’s platform a means for ensuring ‘free speech’?
  2. Should the US authorities, who have clearly recognised the dangers of the technology in Chinese corporations (TikTok’s status and future in the US may depend on the divestment of Chinese interest in the business due to data security concerns), not also apply similar standards to American corporations?
  3. Is control less harmful when exerted by homegrown didacts?

Margaret Atwood’s timely warning in The Handmaid’s Tale can be aptly applied to any group with the power to preach and manipulate and espousing an extremist philosophy and discourse, not only to a religious source.

Can the subaltern speak in the digital space?

In 1988, the scholar Gayatri Chakravorty Spivak wrote an article titled ‘Can the subaltern speak?’ that provides a pertinent lens for the current debate. When discussing arguments by postcolonial scholars to give a ‘voice’ to the Indian ‘subaltern’ (the economically dispossessed), Spivak asked whether the argument-makers did not end up ‘speaking for’ the subaltern?

In the effort to grant the subaltern – a heterogenous group – ‘collective speech’ did the scholars not end up subsuming diversity in unified classifications that represented their own views of the subaltern voice rather than give voice to the subaltern.

In the AI age, we may further ask whether simply allowing a person/group to post on an issue, when perhaps they may have been ‘nudged’ or ‘manipulated’ or been locked in a ‘filter bubble’ of their own preferences by AI applications, is giving them the right to free speech? The global reach of the internet extends to groups that may not have been allowed, or taught, to critically think on, or indeed, have the knowledge, time or other resources to fully access all the points at issue; does simply putting them in the fraught environment of the internet, often hijacked by the loud, and the loudness magnified by AI, still equate to giving them a ‘voice’ in a meaningful sense?

In any argument-making, the balance of power always tilts in favour of those with the knowledge and resources such as time, insider-knowledge, cultural expertise, technical leverage and financial muscle to shape the contours of the argument.

One may agree with Musk that freedom of speech is fundamental to societies, particularly to democracies, and the United Nations should put it high on its agenda to advocate for mechanisms that make it impossible for powerful individuals and corporations to manipulate the behaviour, actions and emotions of millions of people in the guise of safeguarding their rights. In the AI age, a new category of ‘subaltern’ is emerging: those who are dispossessed of the power to hold the Internet Powers to account.

The subaltern groups may be said to be all those who in their different ways are denied the power to access, or use, digital technologies free from the agendas of those who control the technology but are not accountable for its use. A strange alliance may exist amongst all those who cannot access digital technologies, or walk away from them without incurring almost unbearable costs in access to health, work or other fundamentals, if effect, anyone without choices with regard to digital technologies may be classed as a ‘subaltern’. Just as the economically dispossessed had ranks and hierarchies amongst them so do the digital subaltern and all cases of dispossession are not equal. However, it would be fatally complacent not to keep asking the questions in the global North and the global South:

What stands in the way of the tech billionaires in using AI technologies to promote their own agendas?

What safeguards will guarantee that the continuous interaction of the user with the AI technologies of social media platforms is not, vampire-like, sucking the free-speech rights of the ordinary folk to fatten the powers of the platforms themselves?

Who, in reality, speaks when the internet subaltern speaks?

Image credit: https://unsplash.com/photos/a-white-toy-with-a-black-nose-6UDansS-rPI