broadcast media. This was not simply a matter of pragmatics – decentralized networks are inherently more difficult to regulate than centralized ones – but a fundamental philosophical choice. For de Sola Pool, the new networks or electronic communications required a ‘policy of freedom’ and ‘freedom from government’:
The phrase ‘communications policy’ rings oddly in a discussion of freedom from government. But freedom is also a policy. The question it poses is how to reduce the public control of communications in an electronic era. A policy of freedom aims at pluralism of expression rather than at dissemination of preferred ideas. (De Sola Pool, 1983, p. 8)
Such a ‘policy of freedom’ would be enabled by the perceived plasticity of the medium itself. Unlike broadcast technologies, digital media were seen as being fundamentally shaped by their user communities. The internet pioneer Vinton Cerf argued that ‘the Internet can be anything we can imagine and program it to be. It is a most malleable and evolving infrastructure … With such plasticity, it is how we think about the Internet that matters … The Internet evolves in response to forces acting on it’ (Cerf, 1996, p. ix).
This should not be taken as a sign that internet theorists were necessarily conservative or neoliberal. While some authors were clearly in favour of the free market (Rossetto, 1996), others saw the internet’s roots in anti-market movements associated with the New Left in the 1960s. For example, Mark Poster connected the ‘second media age’ of the internet to postmodernism (Poster, 1994), while Howard Rheingold (1994) envisaged a new communitarianism and a revitalization of the public sphere around digital culture.
Tom Streeter (2011) has argued that internet culture has long been strongly imbued with an ethos of romanticism, which can have both capitalist and anticapitalist variants. Capitalist romanticism may be found in the continuing influence of authors such Ayn Rand, the ‘free minds and free markets’ ideology often promulgated in Silicon Valley, and the enduring appeal of the charismatic outsider who challenges the establishment with a new vision. Think of the appeal of Steve Jobs in challenging the dominance of IBM in the 1980s, of the image of a geeky, t-shirt-wearing Mark Zuckerberg taking on the ‘suits’ in films such as The Social Network, or of Elon Musk promoting cryptocurrencies as disruptors of traditional finance. Even when parodied, as in HBO’s comedy Silicon Valley, Silicon Valley preserves its underlying iconography of a hacker culture. But an anticapitalist romanticism, or at least one that is deeply ambivalent about capitalism, has also been prevalent and its ethos is recognizable in movements such as the open source software movement. This movement claims that collaboratively developed software is superior to proprietary forms, as well as in a long-standing hostility to copyright and intellectual property laws, which create ‘information enclosures’ and barriers to social sharing and collaboration (Vaidhyanathan, 2001). Streeter argues that one of the legacies of romanticism is the expectation that the internet should be open and disruptive by its very nature. Streeter also argues for the importance of such ideas, which exist in a complex but relatively autonomous relationship with dominant corporate interests and institutional forms, in the ongoing shaping of digital environments:
One of the most powerful forces maintaining the Internet’s open, anarchic character, in sum, is our memory of all the romantic stories about the Internet; those stories taught us to expect the Internet to be liberating and unpredictable, and that expectation helps keep it that way. The Internet is open, not because of the technology itself or some uniquely democratic potential hidden inside the technology, but because we have narrated it as open and, as a consequence, have embraced and constructed it as open. (Streeter, 2011, p. 175)
Openness as Public Policy: Safe Harbour and the Communications Decency Act 1996
A characteristic feature of the politics surrounding internet regulation was therefore a considerably greater concern with the potential obstruction of free speech rights than with the role that policy was likely to play in shaping the political economy of the internet or the economic structures of the wider society. While in the United States the Telecommunications Act 1996 was adopted by the Clinton administration with relatively little contention – and indeed would become a template for new legislation in other parts of the world, as the UK government’s Communications Act 2003 illustrates (Freedman, 2008) – the associated Communication Decency Act 1996 (CDA), which set restrictions upon the online distribution of ‘offensive’ content, faced far more organized opposition.
Although the CDA was signed into law by President Clinton, the American Civil Liberties Union (ACLU), the Electronic Frontiers Foundation (EFF), and other bodies had the legislation struck down in district court, in June 1996 in Pennsylvania, on the grounds that it was a threat to freedom of speech, which was constitutionally guaranteed under the First Amendment, and could have the consequence of ‘chilling’ lawful speech or conduct. In his concluding remarks striking down the legislation, Judge Dalzell argued for the distinctiveness of the internet as a uniquely speech-enhancing medium: in his view, the internet’s seeming chaos and cacophony of speech were in fact its greatest assets, so that to restrict them in response to concerns about pornography or offensive speech was like ‘burning down the global village to roast the pig’. Rather than viewing the internet as a communications medium, Dalzell described it as ‘a never-ending worldwide conversation’ and claimed that, ‘as the most participatory form of mass speech yet developed, the Internet deserves the highest protection from governmental intrusion’ (Judge Dalzell, quoted in Godwin, 1998, p. 289).
The US approach to internet governance was certainly not the only existing one. China, for example, chose to maintain strict controls over internet content, developing the Great Firewall – a system of monitors designed to regulate the content available to Chinese users and to restrict the access of foreign companies to the domestic market. In other countries such as Singapore and Malaysia a relatively open internet was promised in order to encourage international investment, while laws were maintained over speech that service providers would be ‘encouraged’ to commit to. In Australia, the Broadcasting Services Act 1992 was extended to internet content, thereby giving administrative oversight of online content to the Australian Communications and Media Authority, as a convergent regulator (Australian Law Reform Commission, 2012).
The most enduring legacy of the CDA and of the manner in which it was struck down has been the separation of the legal obligations of companies that primarily engage with the public through the internet from the legal obligations of organizations that are conceived of as media companies, in a more traditional sense. The ‘safe harbor’ provisions enshrined in Section 230 of the CDA have given technology and social media companies considerable latitude over how they manage content hosted on their sites, without the liabilities associated with being publishers or media companies. It enshrines a distinction between the internet and ‘the media’ that has continued to prevail in discourses and policies related to the internet; and these are quite different from the original concerns about whether pornography or radical speech had constitutional protection under US law. The European Union’s e-Commence Directive, adopted by member states in 2000, followed the US model in setting clear limits to the liability of internet intermediaries for the content they host, although in practice the safe harbour provisions have been less clear in their application in Europe than in the United States (Keller, 2018).
The Changing Internet Landscape
From the perspective of the 2020s, the early years of the internet appeared to be marked by a relative absence of regulation. There were certainly restrictions placed upon users’ access to online content, the most common points of intervention being around the internet’s filtering of content that could be deemed to be pornographic, to qualify as hate speech, to promote terrorism, or to infringe copyright (Suzor, 2019; Zittrain, 2002; Zittrain and Palfrey, 2007). But attempts to regulate access to online content typically experienced a significant pushback, at least in liberal democratic societies with a strong civil society. For example, attempts by the Australian federal government to establish mandatory filtering for internet service providers in the years before 2010 were successfully resisted by a diverse coalition of industry and civil society organizations, although the constitutional basis for such restrictions