The OECD.AI Policy Navigator

As artificial intelligence becomes increasingly sophisticated, it is being weaponised to manipulate human perception at unprecedented scale and precision. Deepfake videos, voice cloning, personalised disinformation campaigns and emotion-targeted content are no longer theoretical risks but active threats affecting elections, public trust, organisational security and individual decision-making. Traditional digital literacy approaches are insufficient because they do not address AI's unique capabilities: algorithmic personalisation that traps users in information bubbles, conversational bots that simulate empathy to extract sensitive data and synthetic media so realistic that even experts struggle to detect manipulation.

This educational guide aims to transform passive digital consumers into informed, critical actors capable of recognising and resisting AI-based manipulation. Its primary objectives are to: (1) explain how AI technologies work in accessible language, demystifying complex mechanisms like neural networks, large language models and emotion AI; (2) illustrate real-world manipulation scenarios including spear phishing, microtargeting, deepfakes and artificial social consensus; (3) provide concrete detection and prevention methods for individuals, educators and organisations; and (4) promote a culture of verification, critical thinking and digital responsibility. The guide specifically targets diverse audiences from schoolchildren and parents to public officials and corporate leadership, recognising that digital resilience must be built across all segments of society.

Developed through collaboration between the Cloud Security Alliance Romanian Chapter, the National Cyber Security Directorate (DNSC), Romanian Intelligence Service (SRI) and Romanian Police, through the Crime Research and Prevention Institute, this initiative leverages institutional credibility and technical expertise. The guide is freely accessible in both Romanian and English, with practical resources including fact-checking platforms, deepfake detection tools and educational materials. Initial distribution focuses on government institutions, educational networks and cybersecurity communities, with documented engagement from law enforcement training programmes and corporate security departments.

The initiative is designed for continuous evolution as AI-based threats develop. The guide emphasises five fundamental directions for societal preparedness: AI-era adapted digital education that trains users to recognise algorithmic manipulation; clear regulations to ban or regulate manipulative AI-generated content; algorithmic transparency through independent audits of AI systems; multidisciplinary collaboration between cybersecurity specialists, psychologists, educators, legal experts, and sociologists; and accessible detection and verification.