Ofcom mandates Online Safety Act compliance for all generative AI tools

This regulatory clarification removes the legal ambiguity surrounding synthetic media, placing the burden of safety directly on platform operators rather than AI model developers alone. For the streaming and digital media ecosystem, this means any service integrating conversational AI or user-facing generative tools must now treat AI outputs with the same moderation rigor as traditional uploads. The move signals a shift toward proactive algorithmic accountability, forcing companies to resource content moderation teams specifically for AI detection. Watch for the publication of Ofcom’s final Illegal Harms Codes of Practice in December 2024, which will define the specific technical benchmarks for Ofcom Online Safety Act compliance.

Ofcom's decision to bring generative AI chatbots under the Online Safety Act 2023 places the UK among the first jurisdictions to explicitly apply existing content-safety legislation to AI-generated outputs. The regulator's approach has drawn attention from other national bodies. In May 2026, Google published new documentation aimed at optimizing websites for generative AI features in Search, emphasizing non-commodity content and agent-friendly structures, signaling that platform operators are already adapting to a world where AI-generated content faces heightened scrutiny from both regulators and search intermediaries. The same report noted that Google updated its spam policies to prohibit attempts to manipulate generative AI responses, a move that mirrors Ofcom's insistence that platforms cannot treat AI outputs as outside their moderation responsibilities.

The business implications extend beyond compliance costs into how companies architect their AI-facing services. Akamai introduced AI Brand Presence to help organizations optimize content for AI search and agentic traffic, combining AI-optimized context delivery with a dashboard that reveals which AI models visit a site and what content they consume. Akamai piloted the technology on its own global website, reporting an 85% increase in citations and a 364% surge in brand presence for general searches where the brand was not mentioned by name. For streaming and media companies subject to Ofcom's guidance, tools like these illustrate the operational complexity of managing how AI systems interact with regulated content at scale.

On the technical side, security researchers are beginning to model AI agents as entities requiring risk scoring and identity resolution, a framework that aligns with Ofcom's expectation that platforms assess and mitigate harms from AI-generated material. Elastic Security Labs published research on integrating non-human identities as entities, including AI agents and agentic workloads, noting that these represent some of the fastest-growing concerns across today's attack surface. The work involves linking an AI agent's sessions to existing entities by tracking its service accounts and connected devices, then assessing the risk of its behavior. This approach offers a concrete technical pathway for platforms that must demonstrate to Ofcom that they can identify, monitor, and intervene when AI-generated content poses a risk of harm to users.