Amid the intensifying atrocities perpetrated by Israeli authorities against the Palestinian people, characterized by relentless violence and grave breaches of the four Geneva Conventions, we are compelled to address Meta on the issue of dehumanization of the Palestinian people, especially during times of crisis

We, a collective of Palestinian human rights and civil society organizations, are deeply astonished at how Palestinian identity continues to be dehumanized and silenced by Meta’s digital platforms. Amid the intensifying atrocities perpetrated by Israeli authorities against the Palestinian people, characterized by relentless violence and grave breaches of the Geneva Conventions 1949, we are compelled to address Meta on the issue of dehumanization of the Palestinian people, especially during times of crisis.

Recently, WhatsApp’s AI image generator created emojis of gun-wielding children when prompted with ‘Palestinian,’ and Instagram’s AI-translation model replaced “Palestinianالحمد الله” with “Palestinian Terrorist”. Both cases indicate a persistent issue. Meta’s generative AI tools are creating offensive and biased results that dehumanize Palestinians. This reflects a pervasive problem of AI-driven dehumanization of Palestinians due to bias in training datasets. This is compounded by the distressing reality that Palestinian voices continue to be silenced. Palestinian content remains disproportionately over-moderated on Meta’s platforms. Moreover, Palestinians are relentlessly targeted with racist and inflammatory speech and incitement to violence which has inevitably led to real-world harm.

It frustrates us to witness the persistent challenges faced by Palestinians on Meta’s platforms. Palestinian civil society organizations have raised their voice for years against the rapidly escalating violations of Palestinian digital rights on Meta’s platforms. Last year, the Business for Social Responsibility (BSR) study found that in May 2021, Meta’s actions had an adverse human rights impact on the “rights of Palestinian users to freedom of expression, freedom of assembly, political participation, and non-discrimination, and therefore on the ability of Palestinians to share information and insights about their experiences as they occurred.”

7amleh’s ongoing monitoring and documentation efforts show that on Meta’s platforms Palestinian voices, especially those of journalists and human rights defenders, face significant and disproportionate censorship and limited reachability in times of crisis. This not only restricts freedom of expression but also hinders access to information. Recently, one of the primary reasons Palestinian content has been aggressively over-moderated is that the company “lowered the threshold” on the level of confidence its automated systems require before acting on Arabic/Palestinian content in particular, from 80% to only 25%. 

Additionally, hate speech, incitement, and disinformation continue to proliferate across Meta’s platforms, undermining the safety and dignity of Palestinian voices online. Meta internal documents acknowledged that the Hebrew hostile speech classifiers were not functioning as effectively because they did not have sufficient data to learn from. Since October 7th, The Palestinian Observatory for Palestinian Digital Rights Violations (7or) documented 532 cases of harmful content, which include incitement to violence and hate speech, across Meta’s three platforms.

We demand Meta take immediate action to protect Palestinian digital rights, especially in times of extreme Israeli violence. It is crucial to ensure that human rights are upheld online and to create safe platforms for all. Meta must also guarantee that Palestinian voices, and others advocating for Palestinian human rights worldwide, are no longer stifled or disproportionately moderated. Content moderation policies and practices must be aligned with international human rights law and the United Nations Guiding Principles on Business and Human Rights.

Meta Must:

Stop AI Bias: We demand an immediate and comprehensive audit of all datasets used for AI training in content moderation. We demand an inquiry into the recent events related to Meta’s AI models, requiring full transparency regarding the sources and characteristics of these datasets, to address concerns related to the negative stereotyping of the Palestinian people.

Stop Silencing Palestinian Voices: Meta must ensure transparency on criteria and thresholds to prevent over-moderation that disproportionately impacts Palestinian content. Meta must enhance transparency of policies concerning the classification and moderation of “Dangerous Organizations and Individuals.” Meta must cease implementing measures that restrict the reach of Palestinian users and limit interactions/comments on their posts. Meta must implement newsworthiness exceptions to ensure Palestinian journalists are not silenced. 

Stop Hate Speech and Incitement in Hebrew: Meta must put an end to the spread of hate speech and incitement on its platforms by fully activating the Hebrew hostile and hate speech classifiers. 

In conclusion, Meta must take immediate and decisive action to rectify the deep-rooted bias that continues to plague its platforms, from AI-driven dehumanization to the disproportionate moderation of Palestinian content. The company risks not only losing the trust of the Palestinian people, but also undermining its credibility in the Arab world.