Commercial self-interest and free speech norms. Does anyone want to disagree with or comment on this passage for my final draft terrorism report (footnotes omitted)? Conflicting Values on Internet Regulation 10. As is often the case, commercial self-interest is clear about its objectives (maximum engagement, minimum intervention) but civil society is conflicted, especially concerning the issue of censorship. As I have written elsewhere , the traditional approach of identifying fundamental human rights which may or may not be outweighed by the counterterrorism interest is complicated by (a) the unsuitability of the human rights framework in the online space and (b) the difficulty of understanding trade-offs between restrictions on free expression and the capacity of the internet in its current form. 11.This is even before one considers the implications, whether moral or practical, of free speech standards under the United State of America’s First Amendment. Naïve assimilation of online content with offline expression ignores some fundamentals: •Evidence strongly indicates that, online, the remedy for bad speech online is not more online: the great Justice Brandeis was writing for a pre-online age . The online trend, experience now shows, is to polarisation not synthesis and understanding. •The online domain provides a platform for ordinary users to communicate their content to millions which they would not have in the real world: it combines both “freedom of speech” with unprecedented “freedom of reach” . Criticisms of broadcast regulation (for example, the 9pm watershed) is typically much more muted than of online regulation, even though many online platforms are broadcast facilities (e.g. YouTube). •Moreover the ‘reach’ of online speech is often determined by the platforms (algorithmically) not the user. As a casual user, if I post something on Instagram I have no way of knowing how many people will view or interact with it. Whether my post has ‘reach’ actually depends on the platform’s assessment of whether it will keep people engaged in the commercial interests of the platform– this process could be described as ‘free speech’ plus ‘commercial reach’ . •Falsehood is very powerful online (for example, through deepfakes and, in future, individually tailored AI-driven communications). •Online accountability or responsibility is weaker than offline because of anonymity or pseudonymity. •Increasing amounts of online are synthetic (generated by AI in one way or another). For some, generative AI will have an adverse impact on freedom of expression, so that more content leads to less speech . •The practicalities of removing user-generated content (digital takedown) are different from the implications of gagging live speech (arrest) or destroying books (pulping) in the physical world. 12.Eric Heinze of Queen Mary University has proposed that one cannot necessarily approach live speech in the same way as online speech . Given the factors listed above, I believe he is justified in querying whether classical free speech norms drawn from JS Mill should apply to online communications. 13.A more nuanced understanding of the applicable principles would allow a better standard of debate on online regulation and is directly relevant to weighting of freedom of expression required by the Act . Regrettably, there is little current academic exploration of this topic which is characterised by broad statements of undifferentiated principle.