The EU has introduced regulation to address the threats to democracy and human rights presented by big tech through social media and AI – moves the US has characterised as ‘foreign censorship’. Global Insight assesses the implications for the rule of law.

That social media platforms pose a serious threat to democracy has been well-established for many years. A decade on from the Cambridge Analytica scandal – which saw social media exposed as a powerful tool to manipulate voter behaviour in the context of the 2016 US Presidential election – online misinformation, extremism and election interference still constitute major challenges to democracies around the world and appear to be getting worse.

The advent of AI and its accelerated development present additional and novel challenges. In response, the EU is attempting to strengthen accountability and oversight through increased regulation. In a climate of heightened geopolitical tensions, Member States are waking up to the danger of excessive dependence on US big tech.

The Cambridge Analytica scandal busted the myth of social media as the ‘digital town square’ by exposing how the personal data of millions of Facebook users was utilised to manipulate voter behaviour in major global elections. The scandal led the UK and EU to adopt strict data protection rules in the General Data Protection Regulation (GDPR) and influenced regulations in the bloc on political advertising. More recently, in a climate of growing concern about the influence of online platforms on democratic processes, the EU in 2022 introduced the Digital Services Act (DSA), which requires large online platforms to address risks related to disinformation, illegal content and electoral integrity, as well as provide transparency on advertising and algorithms. As part of the same package of reforms, the EU also introduced the Digital Markets Act, which aims to prevent global tech companies from abusing their market power.

EU legislation, US pushback

The introduction of the rules has ignited fierce criticism from the US administration, which has characterised the laws as ‘foreign censorship’ and ‘taxes on American firms’, threatening tariff retaliation in response. The EU insists that its laws do not discriminate against US companies and instead enhance freedom of expression by forcing social media platforms to be more accountable and transparent regarding their algorithms and the content users consume online.

Peter Alexiadis, a visiting professor of competition law and digital regulation at King’s College London and Editor of the IBA’s Business Law International journal, says that any action taken by the EU Commission under the DSA is exercised with high levels of transparency, adding that the law is underpinned by rule of law principles, including clear rules, due process and judicial oversight.

The proportionality doctrine within the legislation also prevents states from overly restricting the freedom of speech rights of users. ‘One thing that must not be underestimated is the importance of the European appeals process,’ he says. ‘That process is open, transparent and objective, irrespective of the identity of the complainant.’

Since the DSA took effect, the European Commission has mainly focused on enforcing the provisions of the law that cover harms against children and online fraud, appearing to avoid some of the more controversial areas. In January, a month after the EU fined Elon Musk’s social media platform X €120m for breaching its transparency obligations under the DSA, the US imposed a visa ban on former EU Commissioner Thierry Breton, alleging censorship and coercion of US companies.

Experts who spoke to Global Insight are concerned that threats from the US administration might make the European Commission less willing to robustly enforce its laws against US companies. ‘The effectiveness of the legislation will depend on whether and how the EU can assert their policies against increasing backlash and pressure from host countries of big tech,’ says Markus Beham, the Co-Chair of the IBA’s Human Rights Law Committee. X did not respond to a request from Global Insight for comment.

Alexiadis says that, while backlash from the US will be at the forefront of everyone’s minds, the landmark ruling by a US court in March that found Meta and Google liable for building addictive social media platforms that harmed the mental health of a young person should embolden the EU Commission to pursue violations under the law. ‘EU positions under the DSA can no longer be portrayed as being capricious, protectionist or unreasonable, given these developments,’ he says. Meta, which is appealing judgment, told Global Insight that ‘teen health is profoundly complex and cannot be linked to single app’ and that it remains confident in its record of protecting teens online. Google did not respond to a request for comment.

The present danger of the ‘technoligarchy’

The social media ecosystem is dominated by relatively few companies with market valuations exceeding the GDP of many nations. Meta, which owns social media platforms Facebook, Instagram and Threads, is worth over $1tn. Courtney Radsch, Director of the Centre for Journalism and Liberty at the Open Markets Institute in Washington, DC, says that the extreme wealth and power of tech companies is the real threat to modern democracies and the surveillance capitalism business models of the biggest platforms.

‘Social media is not inherently bad for democracy, but the problem is its ownership by a handful of dominant companies that have monetised every aspect of our attention, behaviour and democratic profile,’ she says. ‘That rewards business models that keep people addicted, which leads to polarisation and extremist content because it is more engaging.’

Many European officials have observed the attendance of tech billionaires at US President Donald Trump’s inauguration in 2025 and at various White House events as evidence of the fusion of economic power and influence at the highest levels of the US government. Christoph Schmon, International Policy Director at US-based digital rights group, Electronic Frontier Foundation, says antitrust law has the potential to be ‘the strongest weapon’ against the domination of online spaces because it can force the structure of a business or service to change. ‘If you have only a few companies mirroring the same content and there’s no options for an alternative, the user experience becomes one of extraction,’ Schmon says.

In a major antitrust ruling in January, a US judge found that Google illegally held a monopoly in online search. But the judge refused to grant the request of government lawyers to break up the company and instead imposed less severe data sharing requirements. In 2025, the European Commission fined Google €2.9bn for discriminatory practices in its ad markets and Apple €1.8bn in 2024 for abusing its dominant position in the app store. Google and Apple did not respond to requests from Global Insight for comment.

AI manipulation and deepfakes

The extraction of personal data looks likely to accelerate with the integration of AI into social media platforms to personalise user experiences and optimise advertising. Human rights advocates – including Irene Khan, the UN Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression – have expressed serious concerns. These relate particularly to dominant platforms backsliding on commitments to electoral integrity, risk management and transparency while increasing investment in artificial intelligence tools, which can enhance political campaign outreach by creating deceptive content and amplifying disinformation.

Susie Alegre, a barrister at Garden Court Chambers in London, who is an expert on technology and human rights, says the ability of social media platforms to manipulate people and use personal data to deliver curated information feeds contravenes the right to freedom of thought and opinion. ‘We are being given constant messaging in a way that we can’t control that is very personalised based on our data, our vulnerabilities and our emotional states,’ she says.

In 2025, the UK introduced the Online Safety Act, which has similar aims to the DSA. It seeks to regulate online services and tackle illegal content, with a greater focus on protecting users, particularly children, from harms. The EU and UK are also considering age restrictions on social media, following Australia issuing a ban on people under 16 using certain platforms. However, these age restrictions mainly focus on mental health issues and will do little to reduce adverse democratic impacts. Alegre says that social media regulations should be aimed at protecting all age groups and need to get to the heart of business models, taking out the possibility of targeted manipulation. ‘Although I don’t think age restriction is the answer it’s a sign that we are waking up to how problematic our current information environment is,’ she says.

The use of AI on social media platforms and the impact of deepfakes in influencing voter behaviour in global elections is another major concern. Dozens of tech companies signed an industry-wide pact to address the use of deceptive artificial intelligence in the 2024 elections. Despite this, a global trend of AI-produced content has emerged, including a deepfake of late dictator Suharto posted ahead of elections in Indonesia, and, in April, a campaign photo posted by a British Member of Parliament prompted accusations of AI manipulation. Leading up to the mid-term elections in the US, hundreds of accounts have appeared on social media featuring AI-generated political content.

In 2024, in response to emerging threats, the EU introduced the AI Act. This imposes a risk-based framework and transparency requirements for companies, including the labelling of AI-generated content. Martin Schirmbacher, a member of the IBA Technology Law Committee Advisory Board, says social media companies seeking to comply with the law tend to overly mark content as AI-generated, which means that it isn’t reliable. ‘It is concerning that we might get to a point where we assume everything is fake until its proven its not,’ says Schirmbacher, a partner at HÄRTING Rechtsanwälte in Berlin. ‘This is not so important for entertainment but of course it is in the context of a war zone.’

Yvonne McDermott Rees is a Professor of Law at Queen’s University Belfast and a principal investigator on the TRUE project. Funded by the European Research Council, it explores the impact of deepfakes on trust in user-generated evidence of human rights violations. Information captured by citizens through their personal devices is vital for international criminal trials because investigators are often denied access to the sites of human rights atrocities. ‘Deepfakes pose a fundamental threat to our trust in evidence. We’re really seeing it being weaponised in the current conflict in Iran where all sides are putting forward content that is either generated or edited using AI and people don’t know what to believe anymore,’ she says. President Trump has been criticised for posting AI-generated images on social media, including in relation to the Iran war.

McDermott says that tech companies shouldn’t allow people to make significant amounts of money from posting AI-generated images on their platforms. ‘Recent investigations have shown that some of the accounts behind the AI content are in completely different countries and are not affiliated to any parties behind the conflict but are generating quite a lot of income from it,’ she says.

In 2025, a BBC investigation found that an international network of spammers based in Pakistan were posting AI-generated images of Holocaust victims on Facebook and profiting from amassing billions of views.

In recent years, people are increasingly turning to AI chatbots to get information and seek advice. A study by the London School of Economics published in December found that conversations with AI large language models can influence people’s political opinions, with information-packed arguments proving the most convincing but least accurate. The researchers warned that powerful actors left unchecked could use these persuasive AI systems to shape public opinion and further concentrate their power. Schmon says that data protection rights for users of AI chatbots need to be strictly enforced given logs containing a repository of ‘perhaps your most sensitive information’ could be weaponised by governments, for example, against women or journalists working in crisis zones. ‘If companies take that information to build profiles, and share and sell those profiles, there are also serious questions about the impact this could have on very personal areas like reproductive health or bodily autonomy,’ he says.

In the UK, lawmakers are pushing to amend the Crime and Policing Bill to make it a criminal offence to create a chatbot that produces illegal content, such as material promoting terrorism, violence and sex abuse. Alegre believes these measures are necessary to hold big tech companies to account. ‘Regulatory fines don’t touch the sides whereas I think if you’re looking at some individual and corporate responsibility from a criminal law perspective then that’s different,’ she says.

Ending dependence on US tech

Rising geopolitical tensions between the US and EU have provoked a movement by European lawmakers, academics and business leaders towards digital sovereignty. The Trump administration’s hostile and unpredictable approach to trade and foreign policy have encouraged Europe to think beyond regulation and towards producing alternatives to US big tech.

Martin Hullin, Director of the Network for Technological Resilience and Sovereignty at German non-profit Bertelsmann Stiftung, commissioned a foundational study into a proposed framework for European digital sovereignty that would consist of European-owned digital infrastructure committed to democratic values and the EU’s sustainability goals. ‘Europe is caught up in a US–China rivalry and the dependence on non-European tech, be it cloud services, AI models, semiconductors, makes us vulnerable to geopolitical pressure,’ he says.

More than 80 per cent of Europe’s digital technologies and infrastructures are imported, and 70 per cent of the foundational AI models used globally originate in the US. The dangers of global dependency on US tech were highlighted following sanctions imposed against Karim Khan, the chief prosecutor of the International Criminal Court, which impacted the ability of the court’s staff to access IT systems. Trump’s threats against Europe over Greenland and efforts to weaken EU tech regulation have also alerted European lawmakers to the national security issues relating to the continent’s dependence on US big tech. In addition, the US Cloud Act – which compels US companies, even those based in Europe, to hand over data to US authorities – directly contradicts requirements of data privacy under the GDPR.

In addition to security considerations, digital sovereignty is necessary to protect modern democracies. ‘With democracy, the risk of the current status quo is related to our cognitive sovereignty,’ says Hullin. ‘We are informing ourselves now through digital tools and are confronted with algorithmic recommendation systems that polarise public opinions more than they bridge different positions.’

And the EU is waking up to the risks. In April, the European Commission awarded a €180m tender to four European cloud providers as part of its efforts to strengthen EU digital sovereignty. The Commission said it chose the companies based on their alignment with the Cloud Sovereignty Framework, which includes supply chain transparency, technological openness, security, environmental considerations and compliance with EU laws and values. The Commission also plans to propose the Cloud and AI Development Act, which aims to massively increase EU data centre capacity and digital infrastructure construction. But private investment is still desperately needed. Bertelsmann Stiftung estimates the switch to EU digital infrastructure will take roughly a decade and require investments of around €300bn.

Cristina Caffarra is an antitrust economist and Chair of the EuroStack Initiative, which focuses on industry-led solutions for European tech and has the support of over 400 companies and organisations. Caffarra stresses that digital sovereignty is essential to EU economic growth, noting that the US economy gains €260bn per year in revenue from European purchases of cloud software services. ‘My initiative’s driven towards how we can mobilise European business towards the mission,’ she says.

In the UK, digital rights groups have led calls for a similar movement towards digital sovereignty. London-based Open Rights Group says the UK risks being left behind as Europe attempts to end its reliance on US big tech for the systems that underpin its democracy, economy and national security. Concern is growing fast over UK government contracts with controversial US data analytics company, Palantir, which has allowed its technology to be used by governments – including by US Immigration and Customs Enforcement and in the Israeli military – to track and target individuals. Global Insight approached Palantir for comment, but it did not respond.

‘The UK needs to follow the EU’s lead and develop a digital sovereignty strategy,’ says Jim Killock, Executive Director of Open Rights Group. ‘Public money should be spent on public code that benefits us all, rather than lining the pockets of Big Tech shareholders.’

Alice Johnson is the IBA Multimedia Journalist and can be contacted at alice.johnson@int-bar.org

Header image: www.illustrationx.com/artists/ThiloRothacker