This phase considers the tools used to generate disinformation and when such activity is likely to surge. For the former, interventions should assess how easy it is to produce high-traction content and where barriers to entry could be raised. This question will become increasingly urgent as GenAI products are released on the mass market, with their potential to turbo-charge the scale, affordability, agility, targeting and persuasiveness of disinformation campaigns. For the latter, the focus should be on building collective readiness and ability to anticipate trends, as well as analysing where and why disinformation is proving effective.4.2 Content creation (production)Marketplace Ensure credible information is available on an ongoing basis, with channels tailored to reach people of all backgrounds and profiles. Invest in local journalism, including stronger platforms for marginalized and underrepresented voices. Prevent predatory and/or micro-targeting of users in the online space, for example via the sale of personal data to advertisers. Support the development of “digital public squares”, from social media to forums, in which high-trust information breaks through and systems are better geared for constructive discovery, debate and learning. Supply Make it more expensive and labour-intensive to produce disinformation at scale. This may include stronger guardrails on GenAI tools such as text and image generators, as well as mixed media “deepfake” technology. Strengthen legal frameworks around copyright infringement (for example, the logo of a known media outlet) as well as non-consensual use or impersonation of someone’s image, voice or identity. Platforms should take stronger action against for-profit human content farms and implement stricter recidivism strategies to prevent disinformation networks from rebuilding after removal. Demand Expose the common features of low-quality or low-trust information, including clickbait, content farms, propaganda, advertising and synthetic or manipulated media. Raise awareness of these red flags and champion signals of information integrity (e.g. clearly cited data and images). Embed access to tools that can support critical thinking, lateral reading and verification of sources in real time. Marketplace Mandate rigorous testing of GenAI tools and services before they enter the market, even in open-source models where decentralization can make enforcement challenging. This may include “red team” exercises that simulate the tactics of disinformers, helping to identify vulnerabilities and design more effective guardrails. Given that open-source AI can be modified and deployed by various actors, testing should occur not just before release but also through ongoing scrutiny and adaptation to emerging threats. Ensure transparent risk assessments that balance the intended value of a product (e.g. entertainment, efficiency, innovation, learning) with its potential misuse and the scale of related harm (e.g. automated or dangerous disinformation in response to a given prompt). Rethinking Media Literacy: A New Ecosystem Model for Information Integrity 18

Rethinking Media Literacy 2025