REBECCA MACKINNON

Early this month, mainlanders trying to create blogs on a Microsoft-hosted service using words like "democracy", "freedom" and "human rights" in the title received a rude reminder: "The title must not contain prohibited language, such as profanity. Please type a different title," said a message.

This warning, equating democracy and freedom with profanity, marks a new milestone in the continuing battle for free expression in the mainland - with a western software behemoth openly throwing its weight behind government censors. And this is only the tip of the iceberg.

China is now engaged in an even larger investment in technology that would help control what its citizens can read on the web and what they can express, thanks to assistance from leading western firms.

A recent study reveals an increasingly sophisticated set of mechanisms through which mainland internet users are prevented from accessing material deemed off-limits by the government. See: http://www.opennetinitiative.net/studies/china.

Though government statements emphasise anti-pornography crackdowns, the primary focus of the mainland's filtering system is on political content.

Public security organs and internet service providers employ thousands of people nationwide to monitor everything posted online by ordinary mainlanders and to delete objectionable content.

The key to the filtering system, however, is automated technology - equipment and software courtesy of US companies - that enable service providers to enter hundreds of thousands of banned keywords and web addresses for automatic blocking.

And this automation is where the true power lies. The Cisco routers sold to China have the ability to block not only the main addresses for websites, but also specific sub-pages, while leaving the rest of the site accessible.

The mainland is clearly not sitting on its current success. It has undertaken a massive internet infrastructure upgrade with a US$100-million project called ChinaNet Next Carrying Network, or CN2. Over the next 12 months, new routers will be installed in 200 cities across the mainland. Experts predict the new network will enable the central government to control and monitor online speech even more tightly.

So, are US companies responsible when Beijing deploys their technology to stifle free speech on the internet?

A heated debate continues in the media and in cyberspace. Robert Scoble, a Microsoft employee who maintains a popular technology weblog, rose to his company's defence: "We must comply with the local laws if we want to do business there," he wrote. A few days later Mr Scoble backed down, but Microsoft did not.

Human rights groups want to hold such companies responsible for facilitating restrictive government policies. This spring, US Congressman Christopher Cox reintroduced the Global Internet Freedom Act, legislation that would support the development of technologies to foil attempts by non-democratic governments to "jam", or filter, the internet.

The act says nothing, however, about the role played by American technology companies in the creation of these controls - passively or actively. There needs to be greater public examination of exactly how US technology companies are conducting their business.

There should be consequences for companies found to be deliberately aiding censorship and political repression.

Rebecca MacKinnon is a research fellow at Harvard Law School's Berkman Centre for Internet and Society.