In this presentation I argue that the toxicity of the information environment is not the natural or inevitable result of Web2.0 and its characteristic user participation but has been promoted by social media platforms through their algorithmic design choices. As confirmed by a wealth of empirical evidence, toxicity is profitable. Whilst the EU Digital Services Act 2023 and the UK Online Safety Act 2024 recognise this and seek to intervene through so-called ‘systems’ duties, they leave intact the law that has facilitated the active mismanagement of the online sphere by platforms, i.e. the extraordinary immunities granted to them in the early 2000s by the Electronic Commerce Directive (following s.230 of the US Communications Decency Act 1996). These immunities have shielded, and continue to shield, platforms from liabilities under civil, administrative and criminal law not just for third party wrongdoing on their domains, but for their own wrongdoing – on the basis that they are no more than ‘neutral’ or ‘passive’ pipes through which the content is channelled.