Malak: AI-based multilingual personal assistant to combat misinformation and generative AI safety issues

Misinformation is spreading faster than platforms can handle it: The paper identifies that existing safety measures on social media platforms like content rating and feedback systems are underutilized, particularly by people with lower digital literacy, leaving a significant gap in protection against false and harmful content.

Deepfakes represent a growing threat: The research highlights that AI-generated fake videos and images (deepfakes) are becoming increasingly convincing and easier to create, enabling malicious actors to cause digital, physical, and psychological harm to individuals through social media.

Current safety tools aren't accessible to everyone: Existing content moderation and verification tools require higher digital literacy to use effectively, meaning vulnerable populations—including those with lower education levels or less tech experience—are left more exposed to online harms and misinformation.

A multilingual approach is essential: By designing Malak as a multilingual personal assistant, the researchers recognize that misinformation and AI safety issues are global problems requiring solutions that work across different languages and cultural contexts, not just in English.

User-friendly AI tools can be part of the solution: The paper proposes that an accessible, AI-powered personal assistant can serve as a preventive measure to help users identify misinformation and deepfakes before sharing them, democratizing access to protection tools across literacy levels.