Twitter is, of course, much more than a headline news wire. For many of its 100 million users around the world, it is primarily a source of diversion and occasional amusement. Nor is it alone as the creator of a new kind of global electronic conversation: Google and Facebook, Tumblr and Wordpress, and much of the rest of the global communications industry are among those reinventing the way the world communicates about its daily intrigues, be they prosaic or horrific.

But if the basic purpose (or "use case," as techies like to say) for Facebook is sharing a picture of your kid or "friending" the cute girl in your chemistry class, the use case for Twitter is to get the word out: I have a new job! The police are pepper-spraying us! The big rally is happening downtown at 10:00! Beyonce is pregnant! Steve Jobs is dead!

And you don't even need a computer; just about any cell phone will do.

Like many Internet media companies, Twitter positions itself as a platform - a utility-like entity that provides a set of tools for people to use as they see fit.

Unlike Time Warner or The New York Times or Reuters, Twitter is not a "content creator," to use the vernacular. Rather it is a proud democratizer of content creation, neutral as to the substance of digital bits of information but open to anyone who has something to say. Twitter doesn't report the news; rather, people report or retransmit the news on it.

As the reach of Twitter and the other Internet media companies extends across the globe, though, it's becoming apparent that they are not just enablers of communication, they are publishers, wrestling with classic publishing problems. They make decisions about what types of words and pictures are suitable, they determine how to respond to would-be government censors, they struggle with how to organize information in a useful fashion, and they even worry about how to handle advertising in a way that doesn't alienate customers.

While Google seems a bit New York Times-ian (smart, thorough, reliable, and a little arrogant) and Facebook tends toward People magazine or USA Today (something for everyone, clean and generic, more concerned with the softer side of life), Twitter is their tabloid cousin: loud and freewheeling, light on rules, heavy on sensational hard news, encouraging risk and experimentation.

In spite of that - or perhaps because of it - Twitter has become one of the most important news purveyors of the 21st century.

The December Twitter messages from the Bahraini activist @Nezrad, translated by the website Global Voices, are painful to read: "because of some words I wrote on Twitter, I was arrested, shackled, blindfolded, and interrogated for five hours. I suffered a lot from being shackled, from thirst, insults and standing without getting to sit down." In a series of 20 messages, he tweeted details of his imprisonment and the mistreatment of his fellow prisoners.

In Bahrain, as in many other countries, you tweet at your own risk - and hope that if the government wants your name so that it can kick down your door, Twitter will protect your identity and make sure your words get out to the world uncensored. Just as traditional Western journalism organizations defend their reporters and fight restrictions on press freedom, so Twitter finds itself on the front lines of a global battle over free speech on the Internet.

By positioning themselves as open platforms and, at least in the United States, not responsible for what people post on their services, Internet media companies have tried to avoid many of the knotty issues old-media editors and their lawyers deal with every day. Is a given message or picture protected free speech, or does it fall into a category - defamation, for example, or copyright infringement, or shouting fire in a crowded theater -that enjoys no constitutional protection, or violates another country's laws?

Yet social media services are far from being unregulated forums for free speech. Most ban many types of expression that are legally protected in the United States - pornography, hate speech, and various forms of advertising, for example - and generally reserve the right to block or delete a user's account.

These companies do, however, have very different policies for how they handle different types of information. Facebook requires real names on its accounts, prohibits hate speech, harassment, and pornography, and often works with governments that want to block information that's illegal locally. In Turkey, for example, it's against the law to insult Ataturk, the founder of the country. Therefore, on Facebook in Turkey, you will not find any insults of Ataturk. The company employs hundreds of people to review terms of service violations, field complaints, and block or remove information deemed objectionable.

Twitter takes a more hands-off approach. Alex Macgillivray, the company's general counsel, declared last year that the company was "from the free speech wing of the free speech party," a stance that has been affirmed by the company's CEO, Dick Costolo. Real names are not required on Twitter, and even hate speech and pornography are not banned outright. (A U.S. judge ruled recently that a stream of hateful and threatening anonymous tweets did not constitute illegal cyber-stalking.) Partly because tweets are so brief (it's hard to violate a copyright in 140 characters) and partly because the company is so young and relatively small (no employees on the ground in countries where they could get arrested), Twitter is still able to fl it over an international media landscape mined with onerous restrictions on electronic communication.

That's one reason Twitter has played such a prominent role as an organizing tool and source of information in recent political protests around the world. Repressive governments, caught off-guard by a free-form news channel they can't control, have been reduced to using blunt force. Egypt shut down its Internet almost entirely at the height of the Arab Spring protests (and Twitter, working with Google, responded by creating a system in which people could phone in tweets). China has its "great firewall," which blocks Facebook, Twitter, and Google, among others. In December, Russian agents apparently flooded Twitter with "junk tweets" to drown out news and interfere with online organizing during that country's recent protests.

Industry officials (and their lawyers) say democratic governments can be as big a threat to online press freedom as iron-gloved dictators. One executive recalls hearing French President Nicolas Sarkozy insist that online conversations had to be reined in. "He was mainly concerned with gossip about his wife," said the official, who declined to be identified due to the sensitive nature of the issue.

In India, which prides itself on its democratic ways, the government summoned Internet executives in December and demanded they implement a censorship regime (they declined). The UK government toyed with shutting Twitter down during the London riots last year, though a recent analysis by the Guardian newspaper showed that other users quickly corrected inflammatory tweets during that tumult, and the service was an important channel for sharing safety and clean-up information.

The ugly truth is that no government - even that of the United States - likes a free speech free-for-all. Federal prosecutors frequently seek court orders demanding information on social media users, or insist that a specific bit of content be removed from the services. In many cases these court orders are secret, and the companies quietly turn over the information. Oftentimes, prosecutors are focused on cyber-crimes such as credit card theft, movie piracy or child pornography, but in the post-9/11 era, they increasingly deal with politics and national security.

Twitter, to the delight of free speech advocates, shed a bright light on these secret court orders last year when the government came hunting for information on people associated with WikilLeaks. Twitter got court permission to inform the subjects of the request and thus allow them to mount their own legal challenges. (The litigation is ongoing.) "This is a very important example of willingness to stand up to powerful interests when many others were willing to do the government's bidding" said David Ardia, a law professor at the University of North Carolina and a former attorney at The Washington Post. "That was a courageous stance."

Defending free speech on a global service subject to the laws of every country in which people use it is a daunting task. But even as it explores new ways of customizing and possibly filtering the service to meet the requirements of different countries, Twitter seems to understand that if you're not willing to publish without fear or favor, you're not going to be much of a news service.

Preserving its ability to do what it does - allow anyone, anywhere, to say just about anything to everyone - is one part of Twitter's news challenge. The other part, probably more difficult but decidedly less glamorous, has to do with categorization, navigation, and the reliability of information on the service.

In the old days of media - up till 10 years ago, say - anyone building a general-interest news service spent a lot of time dividing the world into categories. Politics, business, sports, lifestyle, and entertainment would be typical categories, and within those categories there would be subcategories. Interested in news about the San Francisco 49ers? Click on sports, then click on football, then click on pro football, and then find the 49ers.

Twitter, though, has no such hierarchy for organizing topics, and is only in the earliest stages of creating one. Its river of information is structured not around topics but around people: you see news posted by people you "follow," and send your news to people who follow you. Big news gets passed around - or retweeted - incredibly quickly, so if you are active on Twitter, it won't be long before someone you follow tweets about the death of bin Laden. The fact that you might see dozens of tweets about that can be information in itself, what Costolo calls "the roar of the crowd."

But the shortcomings of this structure are obvious, and often maddening. Most of the time, you're interested in some of the things an individual or organization is tweeting, but not all of them. Also, it can be hard to tell what the provenance of a bit of news is, and thus hard to judge its reliability - context is almost always absent, since a tweet is so brief. Nor is there a good way to access news your network does not happen to be tweeting about, or to delve deeper into a topic.

Twitter users took a stab at solving this problem with the hashtag, or # sign. If you mark your tweet with a # and a topic, anyone can find it - and others on the same topic - by searching for that hashtag. It's still a crude system, though. Search #moscow, and you'll get activists talking about demonstrations and news organizations talking about demonstration politics - as well as help-wanted ads, weather reports, and bad jokes. Twitter's search function is improving, but if you want only the latest substantive and authoritative updates on the political situation in Moscow, there isn't an easy way to find them.

Numerous companies hope to have a hand in making Twitter's millions of daily messages more useful - think of the third-party "apps" on Facebook or the iPhone - but Twitter is determined to address the core navigation and category issues itself. A recent redesign has made it much easier to get started, presenting a more Facebook-like home page and suggesting subjects you might be interested in. This latter concept could ultimately hold the key to the service's utility - behind-the-scenes math wizardry that anticipates, based on a range of factors, what sorts of things an individual might find most relevant.

Instead of the old news structure of categories and subcategories, Twitter might have user-driven search and customized, real-time sorting of the information stream. For example, I signal that I'm interested in California politics. Based on my behavior and network connections, Twitter divines what sorts of things relating to California politics I might find most relevant.

Is a tweet a piece of news, a comment, an endorsement, or a call to action? Eventually Twitter will know, and attempt to treat it accordingly. In principle, it should be able to take its fire hose of tweets and filter it into a customizable, coherent news service. Sometimes, that news will be live reports from "citizen" journalists. More often, reports will come from traditional news organizations. Still other times they will come directly from the news subjects themselves.

People who don't give a fig about politics but love movies should be able to have their own movie wire. The same will be true for fans of sports, music, or food. Like a great newspaper, though, Twitter might also show the show-biz maven the occasional crime story, or highlight the political event so big that everyone should know about it. Those who just want to hear the roar of the global crowd could listen to that too.

The technical challenges in all of this are daunting, and until recently, the company could barely keep its core service working reliably, let alone solve giant computer science problems. Twitter's commercial challenges are massive too. It is not a public company, and says it has no intention of becoming one soon, but with a private market valuation of $8 billion, it's under pressure to show that it can make money. (Investors include Saudi prince Al-Waleed bin Talal, who bought a $300 million stake, and Russian internet mogul Yuri Milner, who put in $400 million.) Advertising is the obvious solution, but too many ads could be a turn-off for users. As Twitter looks harder at the bottom line, it will be tempting for the company to avoid troublesome political activists and focus on, say, the many ad-friendly TV stars and sports celebrities who have become big users of the service.

Fortunately for Twitter, solving the relevance problem will also solve the advertising problem. As Internet media guru John Battelle wrote recently: "If Twitter can assign a rank, a bit of context, a 'place in the world' for every tweet as it relates to every other tweet and every account on Twitter, well, it can do the same for every possible advertiser on the planet, as they relate to those tweets, those accounts, and whatever messaging the advertiser might have to offer."

And fortunately for the world, solving the relevance problem could also reinforce Twitter's commitment to deliver real news. Virtually every study of what people are looking for on the Internet puts news at the top of the list. And nobody is better positioned to seize that opportunity than Twitter.

• A Permanent Economic Emergency