San Francisco (AFP)

Tech giants are mobilized against threats to the US presidential election, from Microsoft, which has detected new cyberattacks against campaign teams from China and Russia, to social networks that anticipate disputed results.

For months, Facebook, Twitter, Google and Microsoft have been increasing the announcements of foiled cyberattacks and dismantled propaganda operations, orchestrated from abroad, especially from Russia.

Microsoft thus revealed Thursday that a group of Russian hackers, Strontium, had attacked more than 200 organizations involved in the presidential campaign (parties, consultants, etc.).

Also Thursday, the United States sanctioned a Ukrainian considered to be an "agent of Russia" for having tried to "discredit" Joe Biden, and whose arrangements supposed to incriminate the Democratic candidate in the November 3 election had been relayed by Donald Trump himself.

This "clandestine" operation, "directed by Russia" to "influence the opinions of American voters", was to "culminate before the day of the vote", estimated the head of American diplomacy Mike Pompeo, in a statement.

Companies and authorities are keen to show their investment to avoid a repeat of the scandals of 2016.

Four years ago, the campaign was marked by massive operations of disguised influence, mainly carried out by organizations close to the Kremlin, such as the "Internet Research Agency" (IRA), in favor of Donald Trump.

- Domestic threat -

The platforms also expect "hack-and-leak" tactics, where state-linked entities give pirated information to the media and use networks to spread it.

That's what happened with emails from Hillary Clinton, the Democratic candidate in 2016.

Ten days ago, Facebook again indicated that it had removed a "small network" of individuals associated in the past with the Russian IRA.

But in the past two years, the platforms, working with the FBI and other agencies, have become adept at fake account networks that spread false information, conspiracy theories, hate speeches and hijacked videos.

They are now especially facing an extremely tense national context, between the pandemic and the wave of demonstrations against racism, which regularly lead to violence.

Not to mention a major player, very active on social networks, the tempestuous American president.

Donald Trump regularly questions the reliability of postal voting, a popular and proven method in the United States, and crucial in times of health crisis.

He even recently suggested that his supporters vote twice.

Facebook and Twitter have cracked down on several occasions, pinning briefing notes to such posts, and redirecting the public to verified information.

Greater use of voting by mail also means that the results may take longer than usual to be known and confirmed.

- Caution -

The two Californian groups, as well as YouTube (Google), are preparing for disaster scenarios in case their platforms are used to proclaim results or call to challenge them, for example.

In a Fox News interview in July, the White House tenant declined to say whether he would accept the results that came out of the polls.

Twitter, which banned political ads almost a year ago, said on Thursday it would now remove or "tag" "false or misleading posts", undermining public confidence in the election, including possible tweets claiming victory before the official announcement.

Three types of comments are concerned: messages which "create confusion" about the democratic process or the authorities in charge of the poll, unverified information on so-called fraud, and those which interfere with the count, such as " illegal behavior aimed at preventing a peaceful transition ".

Facebook for its part created an online information center on the election and announced that no new political advertising could be broadcast in the week preceding the election.

On Youtube, Google also tries to display as a priority information deemed reliable.

The algorithm giant also took a new measure on Thursday against disinformation.

The auto-fill function in the Google bar will no longer predict the sequence of sentences with options "which could be interpreted as allegations against candidates or parties", explained Pandu Nayak, a vice president responsible for the search engine .

This also concerns the practical details.

For example, if a user types "vote by ...", Google will no longer suggest "by phone" in the drop-down menu.

© 2020 AFP