In South Korea press freedom including internet is strictly controlled by the government. In 2016 South Korea was placed at its all-time low of 70th by the press freedom index released from Reporters Without Borders (RSF).

At the dawn of the commercial Internet, the United States decided not to regulate (or tax) the new infrastructure to avoid slowing or stopping its growth. This was the right decision for the 1990s, a time when many believed that the nations of the world would all be friends (remember “The End of History”), sharing common values, and that the new technology would unleash only good things.

This strategy succeeded beyond expectation in business, but not in politics. International conflict has intruded into the happy world of cyberspace. The utopian views of the 1990s have been shattered, but the unregulated Internet pokes along as if nothing has changed. We could pretend this is not happening, but other nations have different views.

It is not just China that regulates the Internet. European governments are increasingly concerned with the falsehood, hate speech, and incitements to violence that populate cyberspace (and social media). They are concerned because their citizens increasingly demand protection from jihadists and incipient fascism.

The U.S. position remains largely unchanged from the 1990s: that the private entities assembled in the multistakeholder community could manage this unregulated space for the public good. A decreasing number of people outside of North America accept this anymore, and there are an increasing number of initiatives on privacy, hate speech, data localization, security, and anticompetitiveness that regulate part of the Internet.

Hate speech is difficult. All democracies protect freedom of speech, but all (even the United States) have some limitations on it. You can’t cry “fire” in a crowded theater, but you can tweet it. The United States and its First Amendment are an outlier—the Budapest Convention on Cybercrime had to remove hate speech as a cybercrime from the treaty and put it in a separate attachment because the United States could not sign a treaty that made hate speech illegal. Europeans, who have experienced fascism, communism, and jihad much more closely than Americans, are not as squeamish about suppressing these evils and not as suspicious of their own governments.

Our opponents have adjusted faster to the online environment than we have. The Internet and social media give them unparalleled reach and influence that old-style propaganda could never match. And U.S. companies have been their partners in this. The companies’ argument, until recently, was that they only provide a platform, they are not responsible for what others post, and we would not want them to be censors. Their business models are built on this, which explains requests to be exempted from Federal Election Commission rules that govern political advertisement. This is the 1990s zeitgeist of Northern California geek libertarianism, and if all the world was Northern California, it might work. The blend of First Amendment absolutism and Internet ideology leads to weird outcomes in the United States, as when a leading online freedom group objects to the removal of Russian and Nazi propaganda from social media.

What the Internet creates is an absence of mediation. Think of mediation as a newspaper editor. Editors hold reporters to standards of evidence. They decide what stories should be published and whether they are news or opinion. Or a librarian, who classifies books and decides which are factual and which should go into the fiction section. No one does this on the Internet, where fiction and fact blend easily. In the 1960s, Daniel Patrick Moynihan said that everyone was entitled to their own opinions, but not to their own facts. The Internet has changed that, and we need to find ways to repair the damage to democracy.

What was new in 2016 was Russian use of social media. The Vladimir Putin regime at first feared social media, seeing it as a U.S. plot to undermine Russia, as it had been used in the Arab Spring and the Green Revolutions. China’s rulers share these fears. Their first response was to control social media in their own countries and to see if they could coerce foreign social media companies into obeying national laws. Their second step was to adopt social media, using it against their domestic political opponents. The third step was to take the tools and techniques they had developed to exploit social media against domestic opponents and apply them to foreign audiences.

The Russians are better in their efforts to disrupt or confuse foreign politics than the Chinese, who are hampered by Han nationalism and a turgid ideology. The Russian intelligence services, guided by Putin’s deep hostility for the United States, at first probably did not believe their luck. The United States’ geriatric politics and deep social inequities gave Russia an opening for political action, and the tools the Russians used to exploit this were a mix of new and old.

The Russians are experts at disinformation, stretching back to the czarist services and the “Protocols of the Elders of Zion” or the claim that U.S. military laboratories created AIDS (“Operation Infektion”), fabrications that many still believe are true. Hacking domestic opponents and leaking embarrassing or fraudulent e-mails was standard practice under Putin for years in Russia. Russian military doctrine talked as early as 2010 of sowing confusion and dissent through information operations. Internet trolls lurking on the comment pages of Western media to make pro-Russia comments have also been a standard ploy.

The online governance model of the 1990s is crumbling in the face of these actions. U.S. companies are caught between China and Europe. Both will regulate. China has not been shy about regulation (or blocking U.S. companies entirely) for national advantage. Frustration in Europe is boiling over social media’s connection to terrorism, Russia’s new style warfare, and privacy protection. European officials now say to social media and Internet giants, regulate yourselves (and not just a charade), or we will regulate you.

This means there is still an opportunity for companies to reshape how the Internet functions. Artificial intelligence can provide automated editors, flagging stories and sources as doubtful or at least increasing transparency on sourcing for readers. These programs will at first make errors and flag or remove innocent content, but artificial intelligence programs can be “trained” and can learn from their mistakes and will quickly improve. Companies can create processes to review and remove dangerous content and to provide ways to dispute removal. The key is to bring mediation to the Internet, whether automatic or human.

A recent survey showed that Silicon Valley executives hate regulation and will move to block it, but the rest of the world does not share their views. The United States may have gridlock on how to create mediation, but other countries will move ahead. Internet companies know best how to regulate themselves. They have an opportunity to reduce risk and protect democracy. They should take it.