The intersection of increasing use of Artificial Intelligence (AI) and rising anti-immigrant narratives in the North Atlantic region has led several nations to label AI-generated propaganda as one of the most alarming emerging security threats of our time (The Globe and Mail, 2025). This phenomenon manifests in several ways - narrative warfare, social media bot accounts, disinformation networks, conspiracy theories, and others. These strategies amplify hate content and create a false sense of consensus. In May of this year, the BBC’s UK investigation revealed that several social media accounts promoting far-right nationalism and White supremacy were not based in the United Kingdom (BBC, 2026). Locations varied from Sri Lanka to the United States of America, and many accounts had a history of posting extremist content for both the far-right and far-left factions. Similarly, in January of this year, the White House posted a photograph of activist Nekima Levy Armstrong that had been visually altered using AI. In particular, Armstrong’s skin was darkened, and her facial expressions made her appear distraught (Electronic Frontier Foundation, 2026).
Countries such as Canada and Germany have also experienced a surge in anti-immigrant and far-right political narratives, such as ethno-nationalism, family values, nativism, anti-internationalism, and populism. Both nations experienced significant surges in immigrant and refugee arrivals during 2021–2023, with Canada’s immigrant intake reaching a record high of over 437,000 in 2022, and Germany absorbing more than 1.2 million refugees in 2022 following Ukraine’s invasion (Organization for Economic Co-operation and Development, 2024). This was followed by a sharp decline in 2024–2025 as Canada’s permanent resident admissions dropped to approximately 340,000 and Germany’s asylum applications fell by roughly 40% (Statista, 2025).
Germany
With the rise of far-right nationalism in Germany, several politicians, social media pages, and influencers have circulated AI-generated images and videos to advance their agendas and target specific communities. AfD’s Maximilian Krah shared AI-generated content depicting staged political demonstrations, fabricated scenarios of cultural threats, crowds of immigrants, and Islamophobic narratives (Center for the Study of Organized Hate, 2025). These images and videos garnered over 1 million views across different platforms. Despite the presence of hate speech and xenophobia in this content, no social media platforms took it down, and Krah did not face any legal action. This propaganda eventually proved to be one of the crucial factors in AfD’s wins in some state elections in 2024 (Center for the Study of Organized Hate, 2025). Krah and the AfD achieved key victories through AI-generated propaganda, replacing the need for factual proof with manufactured visuals that confirm the audience’s ideological positions. They reached millions of people quickly and cost-effectively, and avoided accountability for their content by using social media platforms rather than traditional outlets such as TV interviews and press conferences. They took advantage of free speech laws to move their political campaign away from discussing everyday issues and policy measures, using emotional clickbait to take advantage of racial tensions brewing in the nation.
Canada
Canada, like Germany, also experienced a surge in AI-generated anti-immigration content, fuelled by the India-Canada diplomatic crisis of 2023-2025. In 2023, Hardeep Singh Nijjar, a Canadian man of Indian origin, was shot dead, and then Prime Minister Justin Trudeau accused the Indian Government of Nijjar’s killing. India denied the accusations, prompting both nations to recall their diplomats and suspend some visa services (The Hindu, 2025). Ultimately, no further evidence was provided, and by the end of 2025, India-Canada relations improved. However, this crisis led to a sharp increase in anti-immigrant, particularly anti-Indian content online. Multiple social media influencers, bot accounts, and pages began posting AI-generated visuals depicting Indian immigrants as dirty, lazy, illiterate, and taking advantage of the Canadian economy. This propaganda painted immigrants as the scapegoats for rising inflation, housing costs, and unemployment. Further, the Global Project Against Hate and Extremism reported the doubling of crimes, hate speech, extremism, and slurs targeting Indians and other South Asians online (Press Progress, 2024). The Canadian government also approved a record-low number of immigration applications compared to previous years, following the tightening of restrictions and requirements (Statista, 2025). While researching this issue, I noticed that no major Canadian or foreign news outlets had covered it, except for the CBC and Indian news outlets. Despite the prevalence of hate speech and AI propaganda, policy and news discussions addressing this issue remain inadequate.
Analysis & Recommendations
Although the German and Canadian national governments have acknowledged the growing impact of AI-related security threats, no significant actions or policies have been taken to address them (The Globe and Mail, 2025). AI remains a new and ever-changing field for which most legal systems are unprepared. AI-generated content is challenging to stop because it often involves cross-border operations, making legal coordination and enforcement a complex international undertaking. Further, many social media platforms lack comprehensive cooperation and transparency mechanisms, making it hard for governments to pinpoint a person or entity behind an account they can prosecute.
Some recommendations to stop the spread of AI-generated hate speech:
AI Labels: Governments must make it mandatory that all AI-generated visuals posted online bear a visible label identifying them as such. Whenever possible, platforms should include a warning below content that is inaccurate or not supported by factual evidence. To improve digital literacy and enable users to differentiate between synthetic and authentic content, platforms should also include clear information explaining why content has been labelled. This would help strike a balance between freedom of speech and the spread of disinformation. Governments should also partner with technology companies to track AI-generated content that is downloaded or reposted across different social media platforms.
Removal of Hate Speech: Social media platforms must increase transparency and remove hate speech from their platform without extended delays. They must release comprehensive reports outlining their response times, the proportion of decisions reviewed by human moderators, removal rates, and the volume of content removed. Ideally, such audits should be carried out by independent regulators to ensure fairness and accountability across the public and private sectors.
Public-Private Partnerships: Governments and private enterprises must form collaborative frameworks to protect online users through cross-border investigations, content sharing, and financial penalties for failure to remove hateful content. To quickly identify coordinated campaigns, these partnerships should set up safe information-sharing channels among platforms, government agencies, cybersecurity specialists, and civil society groups. To enable prompt action in high-risk situations, such as elections and civil unrest, where AI-generated content could exacerbate instability, cooperative crisis-response procedures should be established.
Political Campaigns: The government should ban political parties and candidates from using synthetic media in their campaigns. Independent oversight bodies must monitor political campaigns in real time for AI-generated content and speech violations before and during elections. Political parties and their members must face consequences for violating terms, including financial penalties and disbarment.
Inter-platform content labelling: Technology companies should collaborate to develop a unified technical standard for AI, ensuring that all AI-generated content is labelled as such across all platforms where it is posted. This would enable identification even after reposting, allowing users to trace the content’s origin across the digital ecosystem. Adoption across the industry would make enforcement easier, as users behind AI-generated content would not be able to avoid prosecution by switching to another platform.
Intelligence and Cybersecurity Measures: National intelligence agencies should implement initiatives to track the individuals behind accounts that post coordinated AI content, using forensic analysis to link IP addresses, device fingerprints, and payment patterns to their physical locations and organizational sponsors. The focus should remain on surveilling coordinated foreign operations rather than ordinary users. Further, cross-border collaboration with intelligence agencies of allied nations can also enhance efforts to understand indicators of compromise, attribution, and other strategies.
Conclusion
Historically, propaganda has been a powerful and effective tool in achieving political aims. Nazi Germany published racist cartoons and images of Jews in the 1940s, as did the USA, targeting the Japanese in the aftermath of World War II (BBC, 2026). Both political strategies led to a wave of xenophobia and mass suffering of the communities being targeted. It is imperative that North Atlantic governments understand the severity of AI-generated content and its wide-reaching impact on their populations. Governments must cooperate with private and government agencies and companies to identify, prepare for, and take down the people and entities behind such content. In the 21st century, online narrative warfare carries as much weight as armed conflict and should be dealt with as such.
Sources
Fife, Robert, and Steven Chase. “Hogue Inquiry Set to Outline Measures Against Election Interference.” The Globe and Mail, January 26, 2025. https://www.theglobeandmail.com/politics/article-hogue-inquiry-final-report-to-propose-measures-against-election/.
Haidar, Suhasini. “Canada Says Its Inquiry Commission Did Not Acquit India on Links to Nijjar Killing.” The Hindu, January 30, 2025. https://www.thehindu.com/news/international/canada-says-its-inquiry-commission-did-not-acquit-india-on-links-to-nijjar-killing/article69157726.ece.
Johal, Rumneek. “Canada’s Far-Right Is Targeting South Asian and Sikh Canadians to Incite Anti-Immigrant Hate.” Press Progress. Accessed July 17, 2026. https://pressprogress.ca/canadas-far-right-is-targeting-south-asian-and-sikh-canadians-to-incite-anti-immigrant-hate/.
OECD. “International Migration Outlook 2024.” Organization for Economic Co-operation and Development, January 15, 2025. https://www.oecd.org/en/publications/2024/11/international-migration-outlook-2024_c6f3e803/full-report/canada_89e5860e.html.
O’Keeffe-Johnston, Paris. “AI-Generated Aesthetics and the Politics of the German Far-Right.” Center for the Study of Organized Hate, October 13, 2025. https://www.csohate.org/2025/10/13/ai-generated-aesthetics-germany/.
Richman, Josh. “Beware: Government Using Image Manipulation for Propaganda.” Electronic Frontier Foundation, January 27, 2026. https://www.eff.org/deeplinks/2026/01/beware-government-using-image-manipulation-propaganda.
Spring, Marianna. “Anti-Immigration AI Videos Traced to Overseas Fakers.” BBC News, May 15, 2026. https://www.bbc.com/news/articles/ckgpyn30dp3o.
Statista. “Refugees from Ukraine in Germany 2024.” Statista Research Department, October 20, 2025. https://www.statista.com/statistics/1333488/refugees-ukraine-germany/.
