Facebook, Google and Twitter reveal little in answers to Senate

The internet giants, which spoke to the Senate Select Intelligence Committee late last year, gave written answers to follow-up questions. We read them so you don't have to.

Left to right: Facebook General Counsel Colin Stretch, Twitter Acting General Counsel Sean Edgett, and Google Law Enforcement and Information Security Director Richard Salgado testified during a hearing before the Senate Intelligence Committee last year.
Chip Somodevilla / Getty Images

They danced. They dodged. They deflected.

A Senate committee on Thursday published answers from Facebook, Google and Twitter to follow-up questions about Russia's interference in the 2016 election, left over from a public hearing late last year. But the 100 pages of responses amounted to little more than recitation and description of the tech giants' terms of service and operating procedures.

"The Twitter Rules prohibit using automation tools for the purpose of generating spam," Twitter wrote in one response.  "User reports are an important signal, and we rely on our community to help identify inauthentic accounts," Facebook wrote in another.  

Twitter additionally said its ability to respond to the senators' questions was limited because it couldn't "comment on whether or not we received requests related to any specific law enforcement investigations."

All those answers aren't likely to please the senators, who already criticized the trio for not doing more to safeguard their collective services from a Russian misinformation campaign. Representatives from each of the senator's offices didn't respond to requests for comment.

The written responses come at a time when the tech industry is increasingly under Washington's microscope. Over the past year, the industry has shifted from a darling of the economy to accidental stooges for a Russian plot to interfere with the presidential election.

Meanwhile, the tech industry has increasingly spoken out against President Donald Trump's actions, from his administration's travel bans and his announcement that transgender people cannot serve in the military, to the Federal Communications Commission's vote to end net neutrality rules that ensured all internet traffic be treated equally.

Questions over collusion

One of the big questions hanging over the investigation into Russian interference in the election has been whether any campaign colluded with the Russian government. Senators tried to find out with their questions what the social media companies might be able to pull from their vast stores of data to shed light on this question. The answer was, not much.

Did the Trump campaign interact with much of the content created by Russia-affiliated accounts? Facebook said it saw "insignificant overlap" between the targeted content used by alleged Russian conspirators and Trump's presidential campaign. 

"Facebook does not believe it is in a position to substantiate or disprove allegations of possible collusion," the company wrote in its responses, but it will provide what information it has for investigators to evaluate.

Money, money, money

Senators asked a variety of questions trying to understand how much money the tech giants earn in advertising revenue.

Google identified $4,700 in ads from Russian-affiliated advertisers. That's below the $100,000 that seemed to be spent on Facebook. Twitter, meanwhile, said it counted roughly $400 from the Internet Research Agency, the Russian-linked troll farm.

In answer to a question from Sen. Kamala Harris, a Democrat from California, Google wrote:

"With respect to Ads revenues associated with this effort, our extensive investigation identified very limited activity on our platforms: we identified two accounts that purchased approximately $4700 of Google ad inventory. We paid less than $35 of revenue to those actors for ads served on their published content; our earnings were a fraction of that amount."

In an interview with CNET, a Google spokeswoman later clarified it paid $35, total, in ad revenue to legitimate advertisers whose ads ran alongside content from malicious actors. Such content may have posted on any platform that runs Google ads. That could potentially include YouTube videos from Russia-affiliated accounts, or posts to Russia-affiliated websites.

Who's scamming and spamming

The senators asked questions about how the companies detect abuse of their platforms. When it comes to fake accounts, Twitter detects approximately 450,000 "suspicious" logins a day that it says may be bots, or computer programs created to automatically post and respond to things on Twitter. Additionally, it appears the problem is getting bigger. Twitter said it "challenged" an average of 4 million suspicious accounts a week in September, more than twice the rate from the same time in 2016.

Twitter also said it isn't worried about false news on its service. "We have observed our users engage with false information by refuting it," the company said. "Notably, those refuting retweets generated significantly greater engagement across the platform compared to the Tweets spreading the misinformation -- 8 times as many impressions, engagement by 10 times as many users, and twice as many replies."

Fake political events

Harris asked Facebook about events posted on its network. In response, Facebook said there were 129 events posted across 13 different pages run by the Internet Research Agency, a Russia-affiliated group involved in spreading misinformation online.  Facebook estimates 338,300 unique accounts saw the events. 

"About 25,800 accounts marked that they were interested in an event, and about 62,500 marked that they were going to an event," Facebook said. But Facebook said it didn't know whether any of the events ever took place in real life.

The 2017 elections

Twitter said it's not aware of any state-sponsored efforts to interfere with American elections in 2017, including the Virginia and New Jersey gubernatorial elections. The only thing the company did note was a "surge in automated followers for a candidate in a recent Senate election," which Twitter said does not appear to have been state-sponsored. Google, by the way, also said it didn't detect anything.

Facebook didn't give a direct response to whether it saw any state-sponsored operations associated with the 2017 American elections. "We have learned from the 2016 election cycle and from elections worldwide this last year," the company wrote. "We have incorporated that learning into our automated systems and human review and have greatly improved in preparation for the upcoming elections. We hope to continue learning and improving through increased industry cooperation and dialogue with law enforcement moving forward."

To 2018 and beyond

Facebook, Google and Twitter all pledged to do more going forward. They've begun working together in a group called the Global Internet Forum to Counter Terrorism, to help them identify and deal with "bad actors" on each of their platforms. Through that program, Twitter said the group has created a database of 40,000 "hashes," or accounts that have violated their respective policies and been removed. Twitter says Microsoft, LinkedIn, Oath and Snap have joined this effort.

Twitter reiterated its promise to launch a "Transparency Center" that it said will be an "industry-leading" effort to give researchers access to its data. One feature, for example, will be better search capabilities. It will also include more detail about who is advertising on Twitter and how those ads are targeted. The company didn't immediately respond to a request from CNET for information on when the transparency center will open.

CNET's Andrew Morse, Terry Collins and Laura Hautala contributed to the report.

First published Jan. 25, 3:17 p.m. PT.
Update, 4:01 p.m.: Adds additional details.
Update, 4:10 p.m.: Adds background details about Silicon Valley's increasing political speech, and details about Twitter and Google's Russian-linked advertising revenue, and Twitter's new "Transparency Center."
Update, 4:27 p.m.: Adds responses from Twitter, Google and Facebook about whether they detected state-sponsored actions during the 2017 American senate and gubernatorial elections.
Update, Jan. 26, 11:34 a.m.: Clarifies Google's response to a question from Sen. Kamala Harris.