A small portion of the ads, which began in the summer of 2015, directly named Republican nominee Donald Trump and Democrat Hillary Clinton, the people said, although they declined to say which candidate the ads favored.
Most of the ads, according to a blog post published late Wednesday by Facebook's chief security officer, Alex Stamos, "appeared to focus on amplifying divisive social and political messages across the ideological spectrum — touching on topics from LGBT matters to race issues to immigration to gun rights."
The acknowledgment by Facebook comes as congressional investigators and special counsel Robert S. Mueller III are probing Russian interference in the U.S. election, including allegations that the Kremlin may have coordinated with the Trump campaign.
The U.S. intelligence community concluded in January that Russia had interfered in the U.S. election to help elect Trump, including by using paid social media trolls to spread fake news intended to influence public opinion.
Even though the ad spending from Russia is tiny relative to overall campaign costs, the report from Facebook that a Russian firm was able to target political messages is likely to fuel pointed questions from investigators about whether the Russians received guidance from people in the United States — a question some Democrats have been asking for months.
Facebook reported in its blog post Wednesday that about one-quarter of the ads in question were "geographically targeted," although company officials declined to provide specifics about what areas or demographic groups were the recipients. Of those targeted ads, the company said, more ran in 2015 than 2016.
Rep. Adam B. Schiff (Calif.), the senior Democrat on the House Intelligence Committee, said Wednesday that the disclosure by Facebook confirmed one of the ways Russia sought to interfere in U.S. politics and serves as a "profound warning to us and others about future elections."
"This is a very significant set of data points produced by Facebook," Schiff said, adding: "Left unanswered in what we received from Facebook — because it is beyond the scope of what they are able to determine — is whether there was any coordination between these social media trolls and the campaign. We have to get to the bottom of that."
The House panel, whose staff investigators heard briefly from Facebook representatives behind closed doors Wednesday, will follow up with Facebook and other social media companies and platforms to see "to what degree they are able to confirm similar metrics," Sciff said.
An official familiar with Facebook's internal investigation said the company does not have the ability to determine whether the ads it sold represented any sort of coordination.
The acknowledgment by Facebook follows months of criticism that the social media company served as a platform for the spread of false information before the November election. In a statement posted days after the election, Facebook founder Mark Zuckerberg promised to explore the issue but said that 99 percent of information found on Facebook is authentic and only "a very small amount" is fake or hoaxes. In December, however, the company announced that it would begin flagging articles that had been deemed false or fake, with the assistance of fact-checking organizations.
Facebook discovered the Russian connection as part of an investigation that began this spring looking at purchasers of politically motivated ads, according to people familiar with the inquiry. It found that 3,300 ads had digital footprints that led to the Russian company.
Facebook teams then discovered 470 suspicious and likely fraudulent Facebook accounts and pages that it believes operated out of Russia, had links to the company and were involved in promoting the ads.
A Facebook official said "there is evidence that some of the accounts are linked to a troll farm in St. Petersburg, referred to as the Internet Research Agency, though we have no way to independently confirm." The official declined to release any of the ads it traced to Russian companies or entities.
"Our data policy and federal law limit our ability to share user data and content, so we won't be releasing any ads," the official said. The official added that the ads "were directed at people on Facebook who had expressed interest in subjects explored on those pages, such as LGBT community, black social issues, the Second Amendment and immigration."
Clint Watts, a former FBI agent who has studied Russian online influence campaigns, said Wednesday that Facebook's report served as "validation" for findings by him and his researchers, who he said had spotted what they believed to be Russians posing as Americans to press political messages on Facebook as early as 2015.
He said his analysis showed that Facebook ads in 2015 were largely concerned with divisive social messages and were used to identify other Facebook users most susceptible to messaging. Those users were then targeted with election-oriented ads in 2016, he said.
"We had these suspicions, but we could never see who was purchasing the accounts," said Watts, a senior fellow at the Foreign Policy Research Institute. "Facebook's being brave. They probably could have buried this, and they did the right thing by coming forward."
Stamos, the Facebook security chief, said the company is committed to continuing to protect the integrity of its site and improve its ability to track fraudulent accounts. He said Facebook has shut down the accounts that remained active.
"We know we have to stay vigilant to keep ahead of people who try to misuse our platform," he said.
This year, Facebook announced technology improvements to detect fake accounts and more recently announced that it would no longer allow Facebook pages to advertise if they have a pattern of sharing false news stories. Over the past few months, Stamos said, the company has also taken action to block fake accounts tied to election meddling in France and Germany.
The Internet Research Agency has received attention in the past for its activity.
In 2013, hackers released internal company documents showing it employed 600 people across Russia. Ex-employees who have gone public with their experiences at the company in Internet postings and in media interviews have said their work entailed creating fake Twitter and Facebook accounts and using them to circulate pro-Kremlin propaganda. They said Internet Research Agency employees, for instance, spread derogatory information about Putin critic Boris Nemtsov in the days after his 2015 murder.
In 2015, the New York Times Magazine reported that social media accounts linked to the Internet Research Agency had launched social media campaigns in the United States, including a sophisticated hoax that spread false news of a chemical leak in Louisiana in 2014, apparently to sow chaos and fear.
In its unclassified report in January, the U.S. intelligence community concluded that the Internet Research Agency's "likely financier" is a "close Putin ally with ties to Russian intelligence."
In May, Time magazine reported that U.S. intelligence officials had discovered evidence that Russian agents had purchased ads on Facebook to target specific populations with propaganda. A Facebook spokesman told the magazine that the company had no evidence of such buys.
Under federal law and Federal Election Commission regulations, both foreign nationals and foreign governments are prohibited from making contributions or spending money to influence a federal, state or local election in the United States. The ban includes independent expenditures made in connection with an election.
Those banned from such spending include foreign citizens, foreign governments, foreign political parties, foreign corporations, foreign associations and foreign partnerships, according to the FEC. (Permanent residents who hold green cards, however, are not considered foreign nationals.) Violators face civil penalties, as well as criminal prosecution, if they are found to have knowingly broken the law.
Andrew Roth, Alice Crites, Matea Gold and Ashley Parker contributed to this report.