Search Climate Insights Directory

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 report analyzes the use of generative AI by the German far right to create tailored social media content, including AI-generated influencers and imagery depicting migrant crimes. It highlights the role of the AfD and far-right community groups in spreading these narratives and notes a significant failure in the enforcement of the EU's Digital Services Act (DSA), as most reported AI-generated far-right content remained unlabeled or unremoved.

The document analyzes the landscape of Israeli far-right extremism, identifying Kahanism as a foundational Jewish supremacist ideology that informs contemporary 'neo-Kahanist' political parties and violent groups. It details the activities of settler extremists, such as the Hilltop Youth, and the role of far-right parties like Otzma Yehudit and Noam within the current governing coalition. The report highlights a post-October 7 surge in anti-Palestinian sentiment, increased settler violence in the West Bank, and the ideological convergence between Israeli far-right extremists and global anti-migrant and anti-Muslim movements.

The document argues that current online safety regulations, such as the EU's Digital Services Act and the UK's Online Safety Act, often prioritise larger platforms based on user numbers, leaving a gap where illegal harms are concentrated on smaller, high-risk services. It highlights the adaptive nature of terrorist networks that migrate between platforms to evade detection and notes that regulators often lack the real-time primary data needed for effective enforcement. The brief recommends a shift toward ecosystem-level interventions, improved data-sharing with non-governmental organisations, and the creation of risk-based prioritisation frameworks that do not rely solely on user thresholds.

The report identifies a highly interlinked ecosystem on Telegram where pro-Kremlin narratives permeate conspiracy and radical communities in Europe. It finds that while pro-Kremlin and conspiracy groups are strongly connected across France, Germany, and Italy, the level of interaction with far-right groups varies by country, being stronger in France and Germany than in Italy. The study highlights the role of Russian state-affiliated channels and influencers, such as Alina Lipp, in bridging ideological divides and amplifying disinformation, particularly regarding the Russia-Ukraine war and COVID-19.

The CCOA Toolkit is a comprehensive guide for combating online antisemitism, focusing on the method of 'deconstruction' to dismantle harmful stereotypes. It analyzes how social media algorithms and business models amplify hate, identifies various forms of antisemitism—including anti-Judaism, Nazi ideology, and secondary antisemitism—and provides educational activities to build critical digital literacy. The document also examines the legal landscape, specifically the EU's Digital Services Act (DSA), and advocates for a public health approach to prevention and trauma-sensitive content moderation.

The document analyzes the spillover of conflict-related online risks—such as disinformation, hate speech, and technology-facilitated gender-based violence—into domestic environments. It argues that while the EU's Digital Services Act (DSA) and UK's Online Safety Act (OSA) were not primarily designed for conflict settings, their systemic risk assessments and crisis protocols can be leveraged to hold platforms accountable. The brief highlights critical gaps, including the erasure of war crime evidence due to over-moderation and the role of engagement-driven algorithms in amplifying harmful content. It recommends structured cross-border cooperation, the development of evidence-preservation protocols, and a shift toward outcome-based metrics in platform risk reporting.

The report identifies a persistent FIMI landscape in Poland, characterized by Russian and Belarusian operations targeting Ukrainian refugees, the EU, and democratic institutions. It highlights the use of coordinated inauthentic behavior, AI-generated content, and content laundering to amplify polarizing narratives. While public resilience and civil society interventions limited the impact, the report notes significant platform failures—particularly on X, Meta, and TikTok—and a regulatory vacuum due to the absence of a permanent Digital Services Coordinator in Poland.

The Institute for Strategic Dialogue (ISD) identified 545 domestic extremist accounts in Canada, analyzing 413,288 posts across platforms like X and Telegram. The study found that while the volume of hate speech decreased by 34% from the previous period, anti-migrant and antisemitic content remained prevalent. Violent speech peaked in January 2025, coinciding with Justin Trudeau's resignation and Donald Trump's rhetoric regarding Canada. Key narratives included misogyny, conspiracy theories surrounding Romana Didulo, and anti-government sentiment. The report also highlights a 'Revolutionary Movementarianism' strategy aimed at blending public political activism with the secret build-up of paramilitary structures.

The report identifies 545 domestic extremist accounts in Canada, analyzing over 413,000 posts. It finds that ethnonationalists and white supremacists are the most prolific, while anti-Muslim extremists achieve the highest engagement. Hate speech decreased by 34% from the previous period, with migrants and Jewish people being the primary targets. The document also introduces a new analysis of violent speech, noting peaks coinciding with political events like Justin Trudeau's resignation and Donald Trump's rhetoric regarding Canada.

The report analyzes 883 posts featuring AI-generated content (AIGC) from German far-right accounts, finding that generative AI is used as a cost-effective complement to existing digital strategies. Key actors include the AfD, 'Neue Rechte' outlets, and AI-generated 'influencers'. The content primarily focuses on anti-migrant narratives, attacks on the 'traffic light' coalition government, and the glorification of an idealized German identity. The study highlights a significant lack of platform compliance with the EU Digital Services Act (DSA) and the AI Act regarding the labeling and removal of misleading AIGC.

The report finds that women candidates in the 2024 French legislative elections, particularly those from left-wing parties, faced widespread online gender-based violence (OGBV) and gendered disinformation on TikTok. This abuse often intersected with racism, classism, and religious discrimination. The analysis highlights TikTok's failure to effectively moderate this content, arguing that the platform's automated tools are insufficient and that its lack of action contradicts obligations under the EU's Digital Services Act (DSA). The report notes that such harassment can lead candidates to reduce their online activity, thereby undermining democratic participation.

The document warns that figures like Michael Flynn and Ivan Raiklin are promoting conspiracy theories and fantasies of 'retribution' against political enemies, which may drive some supporters to commit violent acts if the government does not enact such measures following an election loss. Social media analysis of X users following these figures shows a steady increase in posts containing violent and revenge-oriented keywords, spiking after the July 2024 assassination attempt on Trump. The briefing notes that while QAnon followers were often told to 'trust the plan,' the current rhetoric may leave supporters feeling obligated to take matters into their own hands.

The document identifies a Russian propaganda campaign that exploited the aftermath of hurricanes Milton and Helene to foster domestic discontent in the US. Russian state-affiliated outlets, including RT and Sputnik, amplified narratives claiming the Biden administration diverted disaster relief funds to Ukraine and migrants. The analysis highlights a strategy of amplifying existing divisive US domestic discourse and the use of AI-generated imagery to portray devastation. The report notes that these campaigns spread largely unchecked on social media platforms, particularly X, to erode trust in democratic institutions and reduce support for Ukraine.

The document analyzes how conspiracy groups and extremist actors exploited the aftermath of Hurricane Helene to spread falsehoods about FEMA, target public officials with antisemitic abuse, and incite violence. It highlights the convergence of climate denialism, election conspiracies, and hate speech, while criticizing the failure of X's moderation and Community Notes to curb the spread of harmful content.

The Institute for Strategic Dialogue (ISD) found that platforms including Meta, TikTok, YouTube, and X are failing to honor public commitments to mitigate deceptive AI-generated content (AIGC). Despite the availability of forensic debunks, 154 instances of election-related AIGC were identified that remained unlabeled or unremoved, garnering 51 million impressions. The report highlights specific examples involving Taylor Swift, Kamala Harris, and Joe Biden, and argues that the lack of systematic detection and policy enforcement allows repeat offenders to spread misinformation and undermines public trust.

The Institute for Strategic Dialogue (ISD) identifies systemic failures in TikTok's moderation of hate speech and extremism, noting that 30% of a sampled white supremacist dataset contained multiple policy violations. The report recommends enhancing Trust and Safety resourcing, updating moderator training to recognize nuanced 'dog-whistles', and overhauling the auto-suggest feature which currently suggests workarounds for accessing violative content. Additionally, ISD calls for expanded Research API access for civil society organizations beyond the US and Europe to improve platform transparency.

The report details the manual data collection and coding processes used to identify hate speech and extremist content on TikTok. Because ISD lacked API access, researchers used English keywords, misspelled terms to bypass moderation, and 'snowball sampling' of follower networks to identify 212 videos for analysis. The methodology includes specific definitions for 'far right', 'white supremacy', and 'extremism', and a coding framework to assess whether content violates seven specific TikTok Community Guidelines, including those regarding hate speech, misinformation, and violent organizations.

Analysis by the Institute for Strategic Dialogue (ISD) indicates that the 'specter' of AI is being used to dismiss authentic imagery of the Gaza conflict, deepening confirmation biases and creating a climate of skepticism that facilitates disinformation.

The EU Digital Services Act (DSA), fully in force as of 17 February 2024, establishes a systemic online safety framework for digital intermediary services. It imposes tiered obligations based on societal impact, with the most stringent requirements applied to Very Large Online Platforms and Search Engines (VLOPs and VLOSEs)—those with 45 million or more monthly EU users. These entities must conduct annual risk assessments and implement mitigation measures against systemic risks, including illegal content and negative effects on civic discourse. The DSA introduces transparency mandates for advertising and recommender systems, and establishes a crisis response mechanism for threats to public security or health. Enforcement is shared between the European Commission and national Digital Services Coordinators (DSCs), with potential fines reaching 6 percent of annual turnover.