The founder of defend, Jason Tanner, was listening to BBC Radio 4's Today programme in bed, early in 2025. He  was jolted into action when he heard that Elon Musk, the wealthy owner of social media platform X, formerly Twitter, was commenting as an authority on tragic child gang rapes in the UK. This didn't quite add up with my knowledge of his expertise so out came the laptop and page after page of his platform was read...

The moment everything clicked

After days of watching the same patterns repeat, Jason realised something fundamental:
the idea of “free speech” was being reshaped in front of our eyes.

Platforms claiming to defend it were using algorithms to decide who gets heard and who gets buried. Calls to shut down legacy media were being sold as liberation for the mass consumer, when in fact they reinforced power through communications control towards this platform and away from the established media, now labelled with a very loaded term - the legacy media..

People believed they were speaking freely. But their visibility — their very presence — depended on an algorithm they couldn’t see. Like Jason, they'd heard about algorithms and just assumed they were trendy new things. A touch of the Emperors New Clothes had descended on society, too busy to question or understand. 

That was the moment the campaign took root. Not as a project, but as a responsibility.

Research & Evidence 

Once Jason recognised the pattern, he began documenting it. Not as a researcher in a lab, but as someone watching a system behave in ways that didn’t match its promises.

The more he looked, the clearer the picture became.

Posts challenging dominant narratives were quietly suppressed. Accounts pushing divisive themes appeared in clusters, often created within days of each other. Racialised language spread rapidly, amplified by networks of profiles that behaved more like coordinated units than individuals.

It wasn’t chaos. It was choreography.

The myth of “the only place for free speech”

Jason got the sense that X , under Elon Musk, positioned itself as a last refuge of “real free speech.” It was a seductive message — simple, bold, and repeated endlessly. But the reality was far more controlled.

The platform wasn’t protecting free speech. It was curating it.

The 2025 research breakthrough

In November 2025, Jason published findings that added a new layer to the story. X’s own AI system, Grok, criticised Musk for failing to deploy existing technologies that could detect and remove bots — tools that have been available since 2018.

The implication was stark:
the platform had the ability to clean up its own manipulation, but chose not to.

Bots weren’t a flaw. They were part of the ecosystem.

The  defend campaign for fairer free speech will commission or undertake  independent  volunteer led research in support of fairer free speech whenever resources allow. It will also endeavour to flag-up credible external research it encounters whenever possible and to create a separate snapshot of information that's accessible through internet searches on any given day as it is increasingly aware that information access can be manipulated elsewhere...

12 February 2025
Majority support moderation on social media platforms, global survey shows
A new large-scale, global survey has revealed that most people want harmful social media content such as physical threats and defamation to be restricted. This also applies in the USA where several social media platforms have recently modified their policies in favour of unrestricted free speech. This study, conducted by the University of Oxford and the Technical University of Munich (TUM), has been published this week.

The global debate on whether and how social media content should be regulated has flared up again in recent months. Citing freedom of expression, the platform operators X and Meta have relaxed regulations intended to restrict discriminatory statements. Meanwhile, Australia has banned social media access for children under 16.

There is a widespread perception that negativity is inherent in social media.

Spyros Kosmidis, Associate Professor of Politics at the Department of Politics and International Relations

However, the discussions and research on these issues focus mainly on the attitudes of companies, policy makers and the media. So far little attention has been paid to the opinions of social media users. To address this gap, a team of researchers at the University of Oxford and TUM surveyed around 13,500 people in six European countries and in the USA, Brazil, South Africa, and Australia in the autumn of 2024. In this representative study the respondents answered an extensive questionnaire on the conflicting objectives of freedom of expression and safeguarding against digital abuse and misinformation.

Online safety preferred to unlimited freedom of expression

A clear majority of 79% of the respondents believe that incitements to violence should be removed, with the strongest approval (86%) recorded in Germany, Brazil and Slovakia. A majority of US respondents also agreed with this statement, although to a lesser extent (63%). Only 14% of all respondents believe that threats should remain online so that users can respond to them.

Only 17% think that users should be permitted to post offensive content to criticize certain groups of people. The country with the highest level of support for this stance is the USA (29%) and support is lowest in Brazil (9%). In Germany 15% hold this view.

The respondents were also asked to indicate their preferred vision on a scale between two opposite poles: a social media platform with unlimited freedom of expression and one free of hate or misinformation. In all countries the majority tended towards safety from digital violence and misleading information.

No universal consensus

'Influential entrepreneurs like Mark Zuckerberg and Elon Musk have argued that free speech must take precedence over content moderation in social media. However, the study shows that the majority of people in democracies want platforms that reduce hate speech and abuse. This applies even in the USA, a country with a long-standing commitment to freedom of speech in the broadest sense,' says study leader Yannis Theocharis, Professor of Digital Governance at the Munich School of Politics and Public Policy at TUM and co-leader of the Content Moderation Lab of the TUM Think Tank.

'But the results also show that we do not necessarily have a universal consensus with regard to all specific trade-offs between freedom of expression and moderation. People’s beliefs are strongly dependent on cultural norms, political experiences and legal traditions in the various countries. This makes global regulation more difficult,' adds Spyros Kosmidis, Professor of Politics at Oxford's Department of Politics and International Relations and co-leader of the Content Moderation Lab of the TUM Think Tank.

30% believe that governments are accountable

Differences were also seen in the question of who should be mainly responsible for creating a safe environment in social media. The share of respondents who believe that accountability should rest mainly with platform operators is relatively similar across the investigated countries, ranging between 39% in Germany, the UK and Brazil and 29% in France, South Africa and Greece. However, there were larger differences among those who prefer government responsibility. Compared to 37% of German and French respondents, just 14% favour a government-led approach in Slovakia.

There is a similar variation among countries with regard to the percentage who believe that individual citizens should mainly bear responsibility, ranging from a high of 39% in Sweden to a low of 17% in Germany. Looking at the aggregate picture, 35% of respondents across all countries chose platform operators, 31% chose individual citizens and 30% chose governments as the preferred responsible party.

65% expect abusive comments

But would it be worth regulating the platforms? 59% of those surveyed think that exposure to rudeness, intolerance or hate are unavoidable in social media. 65% expect aggressive comments when expressing their views on platforms. In South Africa, this applies to 81%, in the USA to 73%.

'There is a widespread perception that negativity is inherent in social media', Professor Kosmidis says. 'Even those who might not contribute to that negativity appear to be becoming desensitized to what they read and see. For politicians and policymakers, this is a difficult problem to solve. On one hand, the health of political debate is at stake; on the other, intervention could violate key democratic principles. For what it is worth, users take a clear stance on this trade-off: they prefer some amount of content moderation.'

Nevertheless, a large majority of people believe that the platforms can certainly be venues for civilized discussion. Only 20% of those surveyed say it is sometimes necessary to be rude in social media to get one’s own opinion across.

The study ‘Content Warning: Public Attitudes on Content Moderation and Freedom of Expression’ can be accessed here.

For the representative study, around 13,500 people aged 16–69 in Australia, Brazil, Germany, France, Greece, the UK, Sweden, Slovakia, South Africa and the USA answered questionnaires in October and November 2024. The survey was coordinated by the public opinion research institute Bilendi & Respondi.

File note & Audit: The annoucement for this survey was captured by the Defend UK Campaign for Free Speech on 03/05/2026. The text above is the text that the campaign copied over on that date as a record. The original link that the campaign used is here