Frances Haugen testifies at European Parliament

Wednesday, November 10, 2021

On Monday, Facebook whistleblower Frances Haugen testified in a hearing organised by the Internal Market and Consumer Protection Committee (IMCO) of the European Parliament, aiming to inform the parliament on challenges from social media algorithms and to provide input to the proposed Digital Services Act. Ms. Haugen was invited by MEPs of the committee in October, after she leaked internal Facebook documents (commonly known as the Facebook Files). She has already testified at the US Congress and the UK Parliament.

In the opening remarks, chairs of several other committees acknowledged the impacts of social media algorithms. Raphaël Glucksmann, Chair of the Special Committee on Foreign Interference, claimed social media of being used by foreign powers for interference operations. Drago? Tudorache, who chairs the AI Committee, stressed the importance of developing AI ‘on the basis of rules that protect human rights and interests’.

Ms. Haugen raised concerns on engagement-based ranking algorithms, describing Facebook as ‘a system that amplifies division, extremism and polarisation’. She noted discrepancies in Facebook’s existing content moderation systems, pointing to ethnic tensions in Myanmar and Ethiopia as consequences of such systems failing to cater to languages other than English, particularly minority ones. She said content-focussed solutions are not scalable, expressing support for a ‘content-neutral approach’, and for issues to be resolved through making social media more ‘human-scale’.

Haugen proposed the so-called ‘1-2-3’ risk-assessment model, in which companies evaluate themselves, provide information to the community for input, and ‘articulate’ strategies to auditory bodies. This model would be ‘dynamic’ as community concerns change. On the issue of trade secrets, she said that they ‘can and should not be an excuse for companies to refuse access to data’. She said Facebook is ‘substantially’ less transparent than other major platforms, pointing to Google, Twitter and Yelp as examples. She said Facebook should regularly publish aggregated data for civic society to analyse.

On enforcement, Haugen emphasised the importance of having competent personnel, noting a lack of experts. In response to Alexandra Geese’s question on whether enforcement should be delegated to the member states, she said doing so would not enable a ‘critical mass’. She also suggested requiring names on algorithmic changes as a means toward stronger accountability.

Haugen critised Facebook’s rename to Meta, and expressed privacy concerns regarding the Metaverse. However, she is optimistic on the Metaverse being more ‘human-scale’. Monika Bickert, Meta’s vice president of content policy dismissed the notion of profits and user safety being mutually exclusive. In a blog post, she said the platform is ‘on track to spend more than $5 billion this year alone on safety and security’, with ‘over 40,000 people’ for the job. She reiterated their commitment to improve AI moderation systems, and denied allegations of algorithmic polarisation, citing lack of academic consensus.

[edit]