Rethinking Media Literacy 2025

for transparency, accountability and risk management.14 The absence of global standards underscores the need for a parallel focus on societal resilience. Beyond governance, integrating AI literacy and data literacy into MIL curricula is essential. Currently, most MIL programmes do not include discussions on the political economy of AI-driven business models, despite their profound influence on information ecosystems. Addressing these gaps requires targeted interventions, such as embedding AI literacy in school curricula, training journalists to detect AI-generated content (AIGC) and equipping policy-makers with the necessary tools to assess AI-driven disinformation. The challenge is particularly urgent for younger generations, who are increasingly relying on LLMs for information retrieval and research. Unlike traditional search engines, which encourage lateral reading by presenting multiple sources, LLMs provide a single answer, potentially discouraging critical analysis. This shift has major implications for digital media and information, as students and young professionals may be more inclined to trust AI- generated responses without verifying information through independent sources. Addressing this requires platforms to implement greater transparency in how AIGC is produced, while MIL programmes must adapt to equip individuals with the skills needed to navigate an AI-driven information ecosystem. Examples of emerging solutions include AI-detection tools such as Google DeepMind’s SynthID, which watermarks and identifies AIGC. However, these efforts are still in their early stages and require significant scaling to achieve widespread adoption. Additionally, watermarking techniques have proven to be inconsistent and easily bypassed,15 highlighting the need for more robust and multi-layered solutions. In the absence of robust protection frameworks, any regulatory response to AI must be accompanied by efforts to strengthen public resilience to AI-driven disinformation. This includes proactive MIL interventions, partnerships with fact-checking organizations and collaborations with social media platforms to introduce friction in the sharing of deceptive AIGC. Without such measures, the rapid expansion of AI threatens to accelerate the spread of false information, further complicating an already volatile information landscape. Increasing online harms and risks to digital safety There are also growing concerns about wider online harms, such as hate speech and threats to digital safety, specifically for youth online. Some 78% of youth respondents to a survey conducted by the Office of the UN Secretary-General’s Envoy on Youth reported having experienced digital threats, while 18% experienced them constantly.16Online hate speech has become a pervasive issue, fuelling discrimination, inciting violence and deepening societal divides. The rise of mis- and disinformation, especially during global emergencies, has undermined public trust and stability, demonstrating the global impact of harmful content online. Online harassment, threats and the non-consensual sharing of private information disproportionately target women, creating significant barriers to their participation in digital spaces and public life. Women journalists face attacks that aim to silence their voices, producing a “chilling” effect on freedom of expression.17 Online harassment, abuse and disinformation campaigns are pervasive, often targeting women journalists with gendered threats of physical and sexual violence, and leading to self- censorship, psychological stress and even women leaving the profession. As the US Surgeon General’s Advisory found in 2023: “more research is needed to fully understand the impact of social media on children and adolescents; however, the current body of evidence indicates that while social media may have benefits, there are sufficient indicators that social media can also have a profound risk of harm to the mental health and well-being”.18 The glorification of mass shooters19 and the accessibility of terrorist material on social media fuel radicalization, inspire copycat attacks and amplify violent extremist ideologies, posing significant security and societal risks.20 Recommendation feeds further create the potential to confine users to “echo chambers”, hindering access to diverse sources of information. Nevertheless, exposure to different types of news sources is more likely on social media than in other types of media, and ranking algorithms do not have a significant effect on the ideological balance of news consumption on high-traffic websites such as Facebook or Google.21 However, some algorithmic feeds prioritize posts with high engagement, which can highlight those posts that are more radical and emotionally charged, simply because they receive more engagement.22 This phenomenon is especially dangerous in times of conflict or during elections.23 GenAI, in particular, poses several risks to information integrity, specifically in terms of content creation – AI-generated deepfakes, “hallucinations”/inaccurate information,24 rewriting of historical facts – and content distribution by perpetuating existing societal biases and amplifying discrimination. These risks undermine access to public interest information, which is the cornerstone of democratic societies. Rethinking Media Literacy: A New Ecosystem Model for Information Integrity 9