The International Day to Combat Islamophobia, observed each year on 15 March, should no longer be treated as a ritual of concern followed by political silence. It should be understood as a warning about the changing form of anti-Muslim hostility. The UN General Assembly established the day through Resolution 76/254 in 2022, then strengthened the global response through Resolution 78/264, Measures to Combat Islamophobia, adopted in 2024. In 2025, the UN Secretary-General appointed Miguel Ángel Moratinos as Special Envoy to Combat Islamophobia.
These are important diplomatic steps. But diplomacy alone is no longer enough. Islamophobia is no longer confined to crude slogans, mosque attacks, discriminatory laws, or sensationalist television rhetoric. It is increasingly getting embedded in the digital systems that decide what is visible, what is credible, what is suspicious, and who is marked as risky. The old image of the Muslim as civilisational outsider has not faded with modernity. It is being translated into data, code, and machine logic.
From historical othering to digital reproduction
Anti-Muslim prejudice has always relied on simplification. In early modern Europe, the ‘Turk’ and the ‘Moor’ were cast as threatening opposites against which Christian Europe could define itself. In Spain, Muslims and their descendants remained suspect even after forced conversion; the expulsion of the Moriscos between 1609 and 1614 followed decades of depicting Muslim-origin populations as alien, unreliable, and politically dangerous, as reflected in Britannica’s entry on the Moriscos. The political function of this narrative was clear: if Muslims could be portrayed as permanently unsafe, then exclusion could be justified as necessity.
That logic survived well beyond Europe’s early modern age. It reappeared in colonial administration, Orientalist scholarship, post-9/11 securitisation, and the repeated framing of Muslims as a population to be managed rather than a community to be understood. Today, digital systems have given those older narratives new infrastructure. What once travelled through sermons, pamphlets, and imperial media now circulates through memes, synthetic visuals, manipulated clips, and AI-generated falsehoods. The format has changed; the prejudice has not.
Islamophobia is global, not regional
However, it will be a mistake to speak of Islamophobia as if it were primarily a European issue. Europe, undoubtedly remains a major site of anti-Muslim discrimination, but the architecture of contemporary Islamophobia is global. It stretches from mosque attacks in North America and Oceania to majoritarian propaganda in South Asia, from platform-driven hate campaigns to the algorithmic policing of Muslim identity in authoritarian settings.
The consequences are concrete. On 29 January 2017, a gunman attacked the Islamic Cultural Centre of Quebec City, killing six Muslims and injuring 19 others; Canada now marks the date annually through the National Day of Remembrance of the Quebec City Mosque Attack and Action against Islamophobia. On 15 March 2019, a white supremacist attacked two mosques in Christchurch, New Zealand, killing 51 worshippers during Friday prayers. The official Royal Commission of Inquiry showed how online extremism, institutional failures, and anti-Muslim hatred had intersected.
These attacks remind us that Islamophobia is not just a matter of hurt feelings; it kills. Yet physical violence is only one part of the threat. The other is digital amplification: the ability to manufacture, stylise, and spread anti-Muslim narratives at unprecedented speed and scale.
AI does not erase bias. It can scale it.
AI is sometimes described as neutral because it is statistical. That assumption is intellectually lazy and politically dangerous. Statistical systems learn from social material, and that material is shaped by power, language, omission, distortion, and history.
The US National Institute of Standards and Technology’s Generative AI Profile warns developers to examine completeness, representativeness, subgroup coverage, latent systemic bias, proxy indicators such as language and dialect, and weaknesses in hate-speech filtering. Its broader AI Risk Management Framework makes the same point more broadly: harmful bias is often structural, not accidental.
For Muslims, the implications are profound. If large language models are trained on internet-scale corpora shaped by years of securitised discourse, anti-immigrant agitation, selective war coverage, and sensationalist portrayals of Islam and Muslim-majority societies, then distorted associations can become normalised in output. A system may not use an explicit slur. It may do something more subtle and more damaging: repeatedly associate Muslims with violence, flatten Muslim societies into zones of instability, or frame ordinary Islamic observance as a security signal. This is how historical othering becomes algorithmic bias.
Synthetic hate is already reshaping the information space
This danger is no longer speculative. A 2025 report by the Center for the Study of Organised Hate documented that in India, 1,326 AI-generated hateful posts targeting Muslims across X, Instagram, and Facebook from 297 accounts between May 2023 and May 2025. The report found 27.3 million total engagements. It identified recurring themes including the sexualisation of Muslim women, dehumanising imagery, conspiracy narratives such as “Love Jihad” and “Population Jihad”, and the stylised glorification of anti-Muslim violence. Of 187 posts reported by researchers, only one on X had been removed at the time of publication of the report.
That finding should alarm policymakers across the Muslim world. It shows that anti-Muslim propaganda no longer requires a sophisticated state broadcaster or a central ideological machine. It can be produced cheaply, translated easily, and spread transnationally through engagement-driven platforms. In such an environment, old prejudice does not simply survive. It becomes scalable.
When surveillance meets anti-Muslim suspicion
If biased models are one danger, surveillance systems are another. In a 2025 position paper on AI and counter-terrorism, the UN Special Rapporteur on counter-terrorism and human rights warned that AI tools may be used to aggregate travel histories, social-media activity, family associations, arrest records, and other state-held data to profile risk, despite the fallibility and bias of such systems.
An example is Xinjiang case. In China’s Algorithms of Repression, Human Rights Watch documented how Chinese authorities used the Integrated Joint Operations Platform to facilitate mass surveillance and arbitrary detention of Uyghurs and other Turkic Muslims. The report noted that around 13 million Turkic Muslims in Xinjiang had been subjected to intrusive monitoring and that credible estimates indicated up to one million people were being held in political education camps. According to report, broad and dubious criteria were used to flag individuals as suspicious and generate detention leads.
Xinjiang is a warning to the entire Muslim world. It shows what happens when anti-Muslim suspicion is fused with state data power: prejudice ceases to be rhetoric and becomes infrastructure.
Beyond Pakistan: A Muslim-World Response to Digital Islamophobia
Pakistan has earned diplomatic credibility on this issue. It helped drive both Resolutions 76/254 and 78/264; and in February 2026 its mission to the UN announced that the OIC Core Group on Islamophobia had met the President of the General Assembly to prepare the year’s observance.
But the challenge of Islamophobia in the digital age demands shared standards, shared strategy, and shared institutional imagination. That is why the Digital Cooperation Organization should respond more directly. Its existing work on digital rights and its Ethical AI Governance Toolbox already recognise the importance of inclusion, online harms, data protection, and ethical governance. The next step is obvious: anti-Muslim bias in digital systems should be named as a specific policy risk. Muslim-majority states need common standards for auditing datasets in key regional and national languages as: Arabic, Urdu, Turkish, Persian, Bahasa, French, and English; testing large models for harmful outputs on Islam and Muslim identities; labelling synthetic religious and political content; and establishing procurement safeguards against surveillance systems that can be turned against Muslim populations.
A strategic call to the Muslim world’s tech ecosystem
The struggle against Islamophobia can no longer remain reactive. Muslims cannot afford to appear only as subjects of moderation decisions, targets of surveillance, or objects of datasets built elsewhere. They must become authors of the standards, institutions, and technical norms that govern artificial intelligence.
That means universities must research anti-Muslim bias in datasets and model outputs. Regulators must require transparency and human-rights safeguards. Start-ups must build with cultural and linguistic competence, not imported assumptions. Civil-society groups must pressure platforms to treat anti-Muslim synthetic hate as a governance failure, not a public-relations inconvenience. And regional bodies must stop treating digital governance as a neutral technical field divorced from civilisational dignity.
The real question of March 15
The defining question of this International Day is no longer whether Islamophobia exists. It does. The question is whether it will be engineered into the future by default.
If Muslims do not help shape the norms of AI governance, others will shape them in ways that inherit older prejudices in newer forms. And if that happens, algorithmic systems will not bury old Islamophobia; they will automate it.
The task, then, is larger than protest. It is about power: the power to define standards, to govern systems, and to ensure that the digital future does not once again code Muslim identity as a problem to be managed rather than a civilisation to be respected.
As the battle over artificial intelligence intensifies, the Muslim world must understand a hard truth: in the age of algorithms, dignity will not be defended by rhetoric alone. It will be defended by those who help write the code of power.
Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of ProPakistani. The content is provided for informational purposes only and is not intended as professional advice. ProPakistani does not endorse any products, services, or opinions mentioned in the article.
