43 Fall 2026 One of the most contested aspects of the OSA is its approach to content that is harmful but lawful. While the Act initially proposed broad duties in relation to “legal but harmful” content for adults, these provisions were significantly narrowed during the legislative process. As a result, a substantial proportion of contemporary antisemitism, particularly conspiracy theories, Holocaust distortion and denial, and collective blame narratives, falls outside mandatory takedown requirements. Platforms may choose to address such content through their terms of service, but the Act does not impose a clear legal obligation to do so. Moreover, whilst there is a duty to abide by their own terms, there is no protection against rolling the terms back, and many platforms have greatly reduced their moderation standards since the Act came into force. This creates a structural gap, given that much antisemitic harm arises from cumulative exposure to lawful but toxic narratives rather than from isolated criminal acts. The OSA’s strength lies in its systemic focus. Risk assessments, transparency reporting, and senior management accountability offer tools that could, if rigorously enforced, improve platform responses to antisemitism. The Act also creates regulatory leverage over platform design choices, including recommendation systems that amplify harmful content. There are several limitations. These include reliance on platform discretion, particularly for lawful content; regulatory gaps in the use of VPNs to avoid geo-blocking platforms that are not regulated by the OSA and host extremist content. In addition, there is a reduced regulatory burden for small and niche services, including gaming platforms, despite evidence that such spaces can expose young users to highly radicalizing antisemitic content in practice;26 and an assumption towards human-authored content as opposed to machine generated content offering limited guidance on AI-generated material. These limitations become more acute when antisemitism is generated or amplified by AI systems. UK Attitude Towards Regulating Artificial Intelligence: A Pro- Innovation Approach In its 2023 White Paper,27 the Department for Science, Innovation and Technology prioritized enabling AI growth through flexible, existing regulatory tools rather than prescriptive rules, while acknowledging generative AI's potential for harmful disinformation, and hateful and biased content as risks to mitigate through design principles and guidance. It therefore rejects a dedicated regulatory framework and the imposition of mandatory labelling. This approach reflects a deliberate policy choice to balance risk management with economic competitiveness. In 2025, the UK’s approach to AI regulation28 evolved from its 2023 principles-based framework into a more actionable blueprint, emphasizing its pro-innovation stance through AI Growth Labs – supervised sandboxes that allow businesses to test AI applications in real-world sectors like healthcare, planning, transport, and manufacturing with temporary regulatory relaxations to accelerate adoption and drive economic growth, such as slashing NHS waiting times or expediting planning approvals. This strategy builds upon an established approach among existing regulators without introducing new legislation, incorporating strict safeguards to exclude risks to fundamental 26. Center for Technology and Society, “Hate Is No Game: Harassment and Positive Social Experiences in Online Games 2021” (Sept. 13, 2021), available at https://www.adl.org/resources/report/hate-no-game-harassment-and-positivesocial-experiences-online-games-2021; see also Aoife Gallagher, Ciarán O’Connor, Elise Thomas, Jacob Davey and Pierre Vaux, “Gaming and Extremism: The Extreme Right on Discord” (Aug. 19, 2021), available at https://www. isdglobal.org/isd-publications/gaming-and-extremism-the-extreme-right-on-discord/; see also Garrison Wells, Agnes Romhányi, Jason G. Reitman, Reginald Gardner, Kurt Squire & Constance Steinkuehler, “Right‑wing extremism in mainstream games: A review of the literature. Games and Culture,” 19(4) SAGE 469–492 (2023), available at https:// journals.sagepub.com/doi/10.1177/15554120231167214 27. Department for Science, Innovation and Technology and Office for Artificial Intelligence, “AI regulation: A pro‑innovation approach to AI regulation” (White paper) (Aug. 3, 2023), available at https://www.gov.uk/government/publications/ ai-regulation-a-pro-innovation-approach/white-paper 28. Department for Science, Innovation and Technology and The Rt Hon Liz Kendall MP, “New blueprint for AI regulation could speed up planning approvals, slash NHS waiting times, and drive growth and public trust” (Oct. 21, 2025), available at https://www.gov.uk/government/news/new-blueprint-for-ai-regulation-could-speed-up-planning-approvalsslash-nhs-waiting-times-and-drive-growth-and-public-trust
RkJQdWJsaXNoZXIy MjgzNzA=