Dr. Karlin Lillington:

I have been a technology journalist and columnist for more than 20 years, primarily with The Irish Times. I am grateful to the committee for this opportunity to offer my perspective on these issues.

The existing business model of social media and search platforms, which is based on extracting and monetising as much personal data as possible from users while encouraging them to engage addictively with and return to the platforms, is a foundation for the serious problems we are discussing today. It is a vicious but highly lucrative circle in which clickbait material of hate, outrage, conspiracy and tribalism proves the most engaging, while the micro targeting of ads and content means only a select receptive audience may ever see material that becomes impossible to refute.

Too often, policy discussions focus on the risks posed by social media in established democracies, but the most vulnerable victims are, ironically, those who fight most courageously on behalf of democracy, namely, human rights defenders. For them, online threats can swiftly descend into violence, arrest, torture or death. Activists do not wish to leave the platforms because, despite their serious flaws, they are a major tool of democracy. They allow anonymity, communicate helpful information or help spread irrefutable evidence, and offer easy-to-use encrypted messaging. Many of the proposed solutions and interventions to social media problems, such as the banning of online anonymity or account registrations being tied to formal identity, only exacerbate the problems. If we better understood and more adequately addressed the serious risks and harms to human rights activists, we could better resolve the problems for all of us because studies indicate human rights activists are the outriders for these dangers.

In 2017, Front Line Defenders, an Irish-based international human rights NGO, analysed data on the murders of 312 activists. In 84% of these cases, the activists had received threats, often made online. As the organisation noted, the world's worst regimes know well that attacking the legitimacy and credibility of human rights defenders softens the ground and lessens the reaction when they are arrested, imprisoned or murdered. Women activists are regularly the recipient of some of the most loathsome and sexually explicit threats. In a survey of eight countries carried out last year, Amnesty International found that more than 40% of women who had been abused online feared for their physical safety and 24% feared for their family's safety because online mobs often issue graphic threats against their children.

Facebook has been particularly implicated in human rights reports, ignoring anti-democratic campaigns on the site and inexplicably viewing despots as opportunities to extend platform reach. For example, in the Philippines, Facebook eagerly helped the Duterte campaign learn more about social media use and considered him to be, in its own words, a king of social media engagement, even though his vigilante drug squads were already well-known for carrying out summary executions. The UN harshly criticised the company for its failure for many years to shut down co-ordinated threats that scaled into violence in Myanmar. One Myanmar legislator concluded that Facebook had been dangerous and harmful to the country's democratic transition. Avaaz, a human rights organisation, recently condemned the company for similarly failing to curtail violent threats to vulnerable minorities in Assam. Similar difficulties arise with Twitter, YouTube and any other platform that can carry a message, photo or video.

Human rights organisations have stated that platforms regularly back down in the face of government requests to remove posts by or accounts of journalists, activists and organisations. Acceding to a request of the Indian Government, Twitter recently took down 1 million tweets relating to Kashmir. Activists have been deplatformed in Vietnam. Algorithms that encourage and promote engagement also enable co-ordinated hate and disinformation campaigns such as a fake accounts broadcast propaganda campaign in Sudan to commend military generals who massacred demonstrators last summer.

The platforms need key reforms that will protect pro-democracy activists and start to remedy abuses elsewhere. They should work more extensively with trusted regional and local NGOs to better understand the context of government requests for content and account takedowns. They need to be more aware of and vigorous in assessing possible future harm and interference caused by their actions and promotional programmes. Governments and regulators should foreground risks to activists as they consider ways to manage online problems. Too many proposals undermine grassroots movements toward democracy and are as much a threat to democracy as the online harm they hope to combat. States and regulators will not address these broad problems unless they terminate the core surveillance capital business model of platforms and reduce their operational size to a level such that platforms can begin to manage and fix these problems.