WhatsApp's AI Accused of Bias, Generates Violent Imagery for Palestinian Prompts and Peaceful for Israeli

Meta's new AI outputs conflicting depiction of Palestinian and Israeli characters, fueling ongoing discussions of AI bias amid escalating conflict in the region.

WhatsApp's AI-generated stickers allegedly created images of Palestinian boys holding guns while generating peaceful images for Israeli boys, raising concerns of AI bias amidst the escalating Israel-Hamas conflict.
Prompts containing the words 'Palestinian,' 'Palestine,' or 'Muslim boy Palestinian' returned stickers of boys wearing Islamic garments and appearing to hold AK-47 rifles, while prompts such as 'Israeli boy' or 'Jewish boy Israeli' returned visuals of characters smiling, dancing or playing football.
The AI tool, which combines Meta's text-processing abilities and image-generating capabilities, had previously stirred controversy by producing inappropriate or bizarre imagery during its testing phase.
Meta's internal teams reportedly flagged the issue and it's unclear how long these differences have persisted. Kevin McAlister, a Meta spokesperson, acknowledged the imperfections of the AI models and committed to continuous improvements.
This incident adds to the ongoing criticism of big tech's alleged bias and inability to ensure fair and balanced content, particularly amid global conflicts such as the Israel-Hamas conflict.