Policy Digest #19

The document details regulatory actions, litigation, and legislative updates across several countries focusing on AI, data privacy, and online safety. It also reviews four reports on algorithmic recommender systems, highlighting risks such as the amplification of harmful content and the prioritisation of short-term engagement over long-term user value.

• The Australian eSafety Commissioner published a report on 6 March 2025 detailing how service providers, including Google, Meta, WhatsApp, Reddit, Telegram, and X, address terrorist and violent extremist material and activity (TVE) based on responses to transparency notices issued in March 2024.
• The European Union is implementing several AI-related initiatives, including the launch of the Apply AI Strategy consultation on 9 April 2025, the enforcement of a scientific expert panel on AI as of 30 March 2025, and the publication of a third draft of the General-Purpose Artificial Intelligence Code of Practice on 11 March 2025.
• In the United Kingdom, the illegal content duties of the Online Safety Act took effect on 17 March 2025, requiring platforms to remove illegal content and conduct risk assessments. Additionally, Ofcom fined Fenix International Limited (owner of OnlyFans) GBP 1.05 million on 27 March 2025 for misrepresenting its age verification technology.
• The United States is pursuing several digital policy actions, including the TAKE IT DOWN Act to criminalize non-consensual intimate deepfakes, an FTC public inquiry into platform censorship practices, and a presidential memorandum from 21 February 2025 investigating foreign digital services taxes and fines on American companies.
• An ISD investigation into YouTube's recommendation algorithm found that while age and gender appear to have low weighting in recommendations, this can lead to younger users being exposed to harmful content, including videos on self-harm and suicide.
• Analysis of risk assessment reports under the EU Digital Services Act (DSA) suggests that platforms often treat algorithmic recommender systems as mitigation measures rather than acknowledging them as design-related risks that can amplify harmful content.
• The Knight Georgetown Institute's Expert Working Group argues that platforms prioritise short-term engagement metrics (clicks and likes) for advertising revenue, which can lead to societal polarisation and the spread of low-quality information.

APA

Institute for Strategic Dialogue (2025). Policy Digest #19. https://www.isdglobal.org/wp-content/uploads/2025/05/Policy-Digest-19.pdf

Chicago

Institute for Strategic Dialogue. Policy Digest #19. 2025. https://www.isdglobal.org/wp-content/uploads/2025/05/Policy-Digest-19.pdf.

Wikipedia

{{cite report |author=Institute for Strategic Dialogue |title=Policy Digest #19 |date=23 April 2025 |url=https://www.isdglobal.org/wp-content/uploads/2025/05/Policy-Digest-19.pdf |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{instituteforstrategicdialogue2025policy, author = {{Institute for Strategic Dialogue}}, title = {{Policy Digest \#19}}, institution = {Institute for Strategic Dialogue}, year = {2025}, month = apr, url = {https://www.isdglobal.org/wp-content/uploads/2025/05/Policy-Digest-19.pdf}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }