Resumen ejecutivo en español | Résumé en français
This report shows how Big Tech companies are working to constrain the ability of EU democratic bodies to regulate their activities in the public interest through “trade” agreements, which are binding and permanent.
Digitalization is the defining economic transformation of our time. The benefits to society are well-known, but the harms caused from the expansion of Big Tech’s are still being understood. The EU has started to recognise the urgent need rein in some of Big Tech’s most pernicious practices. The Digital Services Act (DSA), the Digital Markets Act (DMA), along with the Data Act, the Data Governance Act (DGA) and the Artificial Intelligence Act (AI Act) are first steps towards ensuring that the digital sector of the economy operates under the same framework of fair play and the public interest as the rest of the economy.
The same EU that is advancing new laws governing the digital economy is promoting a digital trade policy that contradicts, and would severely constrain, current and future public interest policymaking in the EU and beyond.
Through a number of bilateral and regional trade agreements Big Tech is seeking to maintain a policy environment which favors private control of technological resources and practices, and data, for supernormal profit. Control over data – and in particular, the ability to transfer data across borders – and keeping their algorithms or source codes secret are the top goals of Big Tech in any “digital trade” agreement.
The EU has finalized trade agreements with a dedicated digital trade chapter with Canada, Singapore, Vietnam, Japan, the UK, Mexico, Chile, Mercosur, and New Zealand. And is currently negotiating digital trade chapters with Indonesia, Australia, India, the region of Eastern and Southern Africa (ESA), and plurilaterally in the WTO.
This research analyses the most dangerous clauses included in the EU digital trade agenda (Free flow of data, ban on data localisation and non-disclosure of source code). It identifies 10 reasons why it will be harmful for European society, Europe’s green agenda and democracy at large:
Digital firms have seen their profits soar during the last few years as a result of a sharp increase in cross- border digital activities. Yet the taxes they pay remain extremely low, including in Europe. A company like Uber, for instance, can easily shift “highest value creation” from the country of its operation to a tax haven like Ireland from where the backend software and analytics are shown to be provided. The European Commission already in 2018 proposed to improve unfair taxation for the digital economy. And, in 2021, the EU joined the global tax agreement reached at the OECD. Yet, EU’s efforts to tax Big Tech could be contradicted by its own digital trade policies.
Nearly all EU trade agreements with digital provisions include a ban on customs duties on electronic transmissions (ETs). This means that while importers of products such as cars, watches, and agricultural goods are subject to duties, or trade taxes, if the same good is electronic – as in the case of books, movies, or music – states are prohibited from imposing taxes. A key argument used by defenders of this ban is that it benefits EU digital export SMEs. But large U.S.-based corporations, including Apple (music), Netflix (movies), and Amazon (books) benefit from the moratorium far more than any SMEs in the EU.
And it is not just direct taxes that Big Tech seeks to prevent through trade agreements. A provision banning governments from being able to require a copy of data to be held locally makes it more difficult for governments to assess corporate profit taxes. Tax havens are increasingly used by Big Tech as “data havens” to prevent government access to data that could have tax implications otherwise.
Maintaining a strong public services sector in Europe will require strengthening algorithmic accountability and up-skilling digital knowledge among public workers. It will also require the use of large data sets by the public sector to improve education, health, transportation, water and electricity distribution, and other public services. Digitalization of public services often involves public-private partnerships with Big Tech corporations. If the data collection of the public service, or the provision of the service itself, is privatized, then so is the data. In order to obtain the data to improve public services public services should maintain the right to access and control the data produced through any partnerships with private companies. Under the proposed EU digital trade rules barring states from requiring the localization of data in the Party’s territory for storage or processing, the required disclosure from companies could be challenged under trade agreements.
The landmark legislation of the GDPR published in 2016 set the global standard for the fundamental rights of data privacy and data protection. Recent trade agreements, like the ones with the UK and New Zealand, include a clause that aims to safeguard the protection of personal data and privacy. However, there are serious doubts that the “safeguards” included will indeed protect personal privacy. Subsequent to the publication of the EU-UK TCA, the European Data Protection Supervisor (EDPS) stated that “[…] the TCA creates legal uncertainty about the EU’s position on the protection of personal data in the context of trade agreements and risks creating friction with the EU data protection legal framework”.
There is a growing body of evidence that AI can exacerbate discrimination and cause harm, either through faulty algorithms which “learn” patterns based on past inequities, or by exacerbating inequalities found in data sets used for training. In 2019, the EC published a White Paper on Artificial Intelligence, which recognised that the increasing use of algorithms in Europe poses specific risks in terms of fundamental rights and in particular in terms of equality and non-discrimination. Further, recent studies have shown that source codes and algorithms which are inter-connected and learn from themselves (machine-learning) can lead to many undesired outcomes which include discrimination based on income, color and gender.
But digital trade proposals proscribe states from requiring source code disclosure. They do contain exceptions to allow Digital trade provisions that bar states from being able to require algorithmic transparency or that copies of data be stored locally constrain remedies for these problems.to requesting judicial or regulatory authorities for investigations, and the EU-New Zealand FTA uniquely expands this to include discrimination and bias. But the Conference of the Federal and State Ministers for Equality of Germany “pointed out that, due to the complexity of the matter, it seemed unrealistic that those affected would be able to detect and pursue algorithmic discrimination.” Furthermore, transparency remedies must also be available for affected parties, researchers, critical engineers, advocates, trade union stewards, and the general public – not just for governments. If algorithmic systems might violate fundamental and human rights to be free of discrimination, AI systems should have to be proven not to do so in advance of their deployment – not after harms are suffered.
The EU Green Deal promotes new technological innovation to resolve the world’s climate crisis. But for the entire world to make the necessary transitions, transfers of climate-reducing technology innovations to ensure their global use will be required. Bans on source code disclosure, and other forms of technology transfer, will render the achievement of the Paris Agreement impossible for many countries.
Countries also need tax revenue (for example, from taxing Big Tech) in order to fund their transition. Big Tech’s proposals to limit the ability of states to tax their activities will reduce those needed investments. The hyper-concentrated and data hungry digital economy promoted by Big Tech and the proposed digital trade rules is also radically at odds with the fight against global warming. The digital economy uses 10 percent of the world’s electricity and generates nearly 4 percent of global CO2 emissions, almost twice as much as the civil aviation sector. Sustainable digitalization cannot co-exist with huge digital monopolies pushing for ever more collection, storing and processing of data on a global scale.
European regulators and legislators have become well aware of the negative impacts of Big Tech’s monopoly practices and powers. Europe has engaged in the most extensive enforcement actions to reduce Big Tech’s market dominance to set a level playing field to ensure fair competition, especially for SMEs. But certain provisions in digital trade agreements, in particular the Understanding on Computer and Related Services (UCRS), bans on source code disclosure requirements, interoperability provisions, and bans on local presence requirements, could undermine these efforts.
The UCRS would guarantee digital infrastructure firms have virtually unrestricted access into countries and rights to operate there with very limited regulation. Countries that agree to the EU’s UCRS agree to include market access commitments for “computer systems, programming including source codes and algorithms, maintaining computer systems and software, and processing and storage of data.” But it would also include those yet to be invented. They could not limit the size or scope of a foreign company’s operations. Applying open-ended disciplines which restrict competition policy remedies to all digital services would benefit the monopolistic practices of Big Tech.
Anti-competitive practices using algorithms are ubiquitous in the online retail sector, where companies like Amazon ensure that their search algorithms privilege their own products or services above those of others. The exceptions included in digital trade rules will not be enough to curve those practices. Those rules still require a suspicion, as they relate to specific cases, and cannot require disclosure as a general rule– individuals must know that they are being harmed and have a suspicion that it is because of the algorithm and convince the regulatory agency.
In 2021, 99.8 percent of all enterprises in the EU-27 non-financial business sector (NFBS) were SMEs. They employed 83 million people. The vast majority of EU-based SMEs that sell online use Big Tech online platforms to reach consumers. SMEs are dependent on platforms’ algorithms in terms of how their products are ranked in search results or are otherwise advertised. Businesses using Big Tech platforms do not have access to the data on their own customers and resulting from their activity on the gatekeeper’s platform, making it impossible for them to compete in a fair market – while the Big Tech platform can use such data for its own business purposes. Digital trade provisions that bar states from being able to require algorithmic transparency or that copies of data be stored locally constrain remedies for these problems.
Furthermore, European proposals in trade agreements propose to fully liberalized the market access for computer and related services so digital infrastructure firms have virtually unrestricted access into countries and rights to operate with very limited regulation. While some may see an opportunity to gain access to foreign markets for European firms, the first-mover and scale advantages of U.S.-based Big Tech means they would likely consolidate their dominance rather than SMEs. In that context, it is difficult to see any scope for protecting or supporting European SMEs.
Europe’s digital industrialisation strategy relies on improving access to data, developing technology and infrastructure, and appropriate regulation. However, the digital trade strategy clashes with Europe’s aims. A great amount of data that is generated in Europe is held by foreign-based companies. European drivers and riders produce data for Uber, European consumers make purchasing choices on Amazon, which the U.S.-based corporation then uses for its own marketing strategies. Digital rules would prevent governments from requiring companies to share this data or requiring data to be held locally. As a result, Europe’s ability to access the large troves of data required to scale digital industrialization will be compromised.
The creation of digital infrastructures, in particular datacenters used for cloud computing is key for Europe’s digital industrialization strategy. Currently, U.S.-based companies now control nearly 72 percent of the European cloud storage market. France and Germany have promoted local data-center infrastructure and the EU proposed the creation of an European cloud, Gaia-X. But the EU’s digital trade rules against data localization proscribe states from being able to require the use of computing facilities or network elements in the Party’s territory for storage or processing. If the EU could not ensure that EU- based data infrastructure is utilized, then cloud carriers such as Amazon, Google, and Microsoft will pursue their data storage and processing needs in cheaper data havens, not in Europe.
Preserving policy space for regulation is crucial to ensuring widespread benefits from digitalization and guaranteeing European fundamental rights in the digital sphere. The digital trade rules are broad and all-encompassing. Public interest regulation would be subject to challenges with only the narrow window of limited exceptions. Future-proofing the ability to regulate according to evolving political and economic landscapes is crucial.
For example, digital trade rules could affect financial regulation and cybersecurity. Decisions in the financial sector are increasingly determined by algorithms which must be subject to regulatory oversight and public scrutiny. Decisions such as who will get a loan for a house or who will be awarded insurance based on credit risks, are increasingly made by data and algorithms. Also, the growing automation of stock markets operations pose enormous risks in terms of financial stability. Despite exceptions for prudential measures, trade provisions bar governments from requiring disclosure of source code in order to ascertain the security of the financial sector and would preclude the regulatory oversight necessary to guarantee financial security.
The Internet of things (IoT) market for digitally connected devices is an emerging concern for cybersecurity specialists. European governments are increasing cybersecurity legislation on IoT devices in order to protect sensitive consumer (including financial) data and safety. Cybersecurity regulation will require standards such as two factor authentication (TFA), and the disclosure of source code to evaluate high-risk algorithms and cybersecurity measures. But the provisions of digital trade rules promoted by the EU would bar states from being able to require necessary disclosure of source code. The exceptions– including in the most recent EU-NZ FTA – still far short of the enormity of the urgent need for more public oversight.
Digital trade proposals in trade agreements represent an effort by Big Tech to further consolidate that upward distribution of income from labor to capital. In discussions on the future of work, the emphasis on job retraining and skill-based technological growth can be useful but should not be a distraction. The most important aspect in shaping who will benefit from expanded technological use will be the policy environment in which that technology is utilized. If workers are not guaranteed their fundamental rights, freedom and autonomy in digitalised workplaces, and if workers do not have a governance stake in the data produced by workers, and instead this data is allowed to be “owned” by the collecting corporation, it will permanently skew the balance of power in further favor of corporations. Whether workers should have economic rights to the data they help produce is a subject being debated. Locking data related commitments under trade agreement will make any such thing impossible, likely leading to a permanent suppression of labour’s collective bargaining power in a digital age.
Big Tech applies extensive political pressure in Europe, and it appears that their lobby activities have resulted in a deregulatory trade agenda that primarily benefits Silicon Valley.
The thinking that more digital trade means that there must be rules governing this trade is misplaced. Trade agreements inherently limit states’ rights to regulate economic behavior. Yet, governments should have the space to advance regulations to ensure human and fundamental rights in the digital economy; promote the use of data and digitalization for the public good; and promote digital industrialization. The EU must ensure that its digital trade agreements support stronger regulation of Big Tech to protect workers, consumers, SMEs, minorities, sustainability, and fundamental rights in the digital sphere.
Click here to read the full report in English.