The U.S. Spies Who Sound the Alarm About Election Interference

The Intelligence Community Campus-Bethesda, a vast office complex covered in vertical panels of maroon siding and mirrored glass, sits on a cliff overlooking the Potomac, surrounded by a forty-acre lawn and a tall wrought-iron fence. Roughly three thousand employees of various United States spy agencies work there. About two dozen of them are assigned to the Foreign Malign Influence Center—the command hub of the battle to protect the Presidential election from manipulation by foreign powers. The center, which opened in 2022, is responsible for deciphering, and defeating, surreptitious efforts to rig or tilt the American vote. The October before an election is the busy season.

Jessica Brandt, a forty-year-old newcomer to the intelligence world, is the center’s first director. Before her appointment, last year, she’d spent her career writing research papers at Washington think tanks, most recently on “digital authoritarianism”—the way dictators use technology to control or manipulate people, at home and abroad. At a thirty-seat conference table in the center, we talked about her move from theory to practice. Now that Brandt has access to classified intelligence, she knows as much as anyone about how foreign powers are trying to tamper with American elections. But she has also experienced firsthand how the polarization of U.S. politics is making it harder to protect the fairness and credibility of the vote. These days, a warning from the U.S. intelligence agencies is no longer accepted at face value. It’s immediately spun for partisan advantage.

Intelligence officials use the term “election interference” to describe attacks on the actual mechanics of vote counting. This is now considered an extremely slight risk. The hodgepodge of state voting systems makes a mass hacking impossible, and recent security upgrades have insured the preservation of paper backups for almost every ballot. The more realistic danger is what officials call “malign foreign influence”: hacks and leaks, bots and trolls, hidden payments and targeted attack ads. Adversaries can use these underhanded tactics to twist public opinion, discredit the vote, and sway its outcome. The center’s job is to mitigate the effects of such machinations, and one of its main tools is forewarning voters through public bulletins.

Yet ever since July 28, 2016, when the director of the C.I.A. began briefing President Barack Obama on the Kremlin’s plot to help elect Donald Trump, it has been agonizingly clear that government alarms about hidden meddling by foreign hands might themselves be perceived as tainting the electoral process. Obama decided not to alert the public before Election Day about the full extent of the Russian conspiracy to assist Trump, fearing that such a disclosure would look like a thumb on the scale in favor of Trump’s opponent, Hillary Clinton, and potentially undermine her widely expected victory.

That, it turned out, was the wrong worry. When the Kremlin brazenly pulled off another hack-and-leak operation the next year, in Europe, France’s response provided an instructive contrast. The Russians had stolen gigabytes of e-mails and other data from the Presidential campaign of Emmanuel Macron. But, before the day of the vote, credibly nonpartisan government agencies informed citizens of a foreign cyberattack; an electoral commission instructed news organizations not to report on the leaked material. David Salvo, the director of the Alliance for Securing Democracy, at the German Marshall Fund, told me that the French government’s action, and the public’s trusting response, was “the best-case scenario.”

The U.S. intelligence agencies, though, waited until two months after Trump won the 2016 election to lay out the sweeping scale of the Russian operation. Instead of averting a partisan battle, the delay ignited one. Democrats argued that the Kremlin’s support rendered Trump an illegitimate leader; Trump and his allies claimed that the intelligence agencies were part of a deep-state conspiracy against him. Seven years later, the fight continues.

Now another U.S. Presidential election may hinge on tens of thousands of votes across a handful of states. Almost any illicit advantage could arguably decide the outcome (and cast doubt on the results), making the race a prime opportunity for foreign meddling. Indeed, intelligence officials and tech-company analysts say that more foreign spies than ever are getting into the game. Clint Watts, the manager of Microsoft’s Threat Analysis Center, told me that the Kremlin’s success in 2016 “convinced almost every authoritarian nation that they needed to jump into this.” And the biggest players, Russia and Iran, are working even harder at election influence than they did in 2016 or 2020. Yet the government’s warnings about foreign schemes are frequently undercut by the efforts of both Democrats and Republicans to weaponize such intelligence. In 2024, Democrats have railed about Vladimir Putin “rooting for” Trump, while Republicans have insisted that Biden-appointed intelligence officials are underplaying Iran’s schemes to defeat the former President—including by plotting his assassination. Representative Mike Turner, an Ohio Republican who chairs the House Intelligence Committee—and who recently put out a statement under the headline “Is the Biden-Harris Administration Colluding with Iran?”—told me, “You don’t hear a lot from the Administration about the malign influence of Iran in hacking the Trump campaign and attempting to kill Donald Trump.” According to people involved in a recent classified briefing on election security, the two sides of the House Intelligence Committee got into a shouting match over the relative scale of the threats.

Brandt told me wearily that she’d heard “the critiques,” and insisted that the center nevertheless stayed focussed on building “the most accurate threat picture we can.” But Kathleen Hall Jamieson, an expert on public opinion at the University of Pennsylvania and the author of a book documenting the effects of the Kremlin’s influence operation in 2016, told me that, with so much partisan noise threatening to drown out the center’s warnings, “our system is still defective.”

A parliamentary election in Slovakia last September marked the advent of a new era in election chicanery. A pro-Russia faction promising to end support for Ukraine was locked in a tight race against a Western-friendly party, Progressive Slovakia. Three days before the vote, an anonymous Instagram account uploaded a recording of the voice of Progressive Slovakia’s leader, Michal Šimečka, describing a “secret plan” to curb alcoholism: raising the price of beer “by seventy per cent to a hundred per cent.” As that recording raced across Slovakian social media, a second one appeared to catch Šimečka conspiring with one of the country’s best-known investigative journalists, Monika Tódová. “Again, will someone walk in and insert the ballots directly?” Tódová’s voice asked.

Šimečka: “This has been taken care of already.”

Tódová: “All right, then. What about me? Is it true that ‘by coincidence’ I will win some kind of valuable prize?”

Šimečka and Tódová called the recordings fraudulent. But while tech-company fact checkers were struggling to determine their authenticity, the Slovakian media entered a legally required forty-eight-hour news blackout before the vote. By the time the recordings were debunked as A.I.-generated deepfakes, the pro-Russia party had won a narrow victory.

The impact of the deepfakes is difficult to quantify. Their exposure did not stop an ally of the pro-Russia party from winning the Presidency the following year. Yet the Slovakian election put Washington on guard that A.I. could blur the boundaries of political reality as never before. This year, U.S. intelligence agencies said that China was probably behind videos of A.I.-generated Taiwanese newscasters reading aloud from a made-up book containing made-up scandals about Taiwan’s President. Brandt, of the Foreign Malign Influence Center, told me that deepfakes “can come in thirty-six thousand flavors,” so teams of forensics experts from throughout the government had conducted a “summer of exercises,” rehearsing plans to quickly evaluate the authenticity and origin of inflammatory material that might surface in the final days of an American Presidential campaign.

In a one-page “election security update” issued in September, the intelligence agencies declared that various foreign adversaries had already posted numerous deepfakes on the Internet. Russia had deployed the most, spreading “conspiratorial narratives” and amplifying “divisive U.S. issues such as immigration” in order to help Trump and hurt the Democrats. Iran had used A.I. “to help generate social media posts and write inauthentic news articles” about everything from the Presidential race to the Israel-Palestine conflict. China was “using A.I. in broader influence operations” but “not for any specific operations targeting U.S. election outcomes.”

Intelligence officials said that, so far, foreign adversaries’ A.I. trickery was “a malign influence accelerant” but not “revolutionary,” in part because those countries had not yet caught up with Silicon Valley in their use of the technology. The report noted that one of Russia’s most widely circulated fakes—a video of a woman in a wheelchair claiming that Kamala Harris had disabled her in a hit-and-run accident—had actually been staged the old-fashioned way, with real actors.

Mark Warner, a Virginia Democrat and the chairman of the Senate Intelligence Committee, told me, “A.I. is the dog that hasn’t barked—yet.” Warner, whose committee compiled a thirteen-hundred-page report on the Russian intervention in the 2016 election, believes that the U.S. is less prepared than ever to fend off foreign influence schemes. Major social-media companies, he told me, have slackened their crackdowns on misinformation—partly because of lawsuits claiming that the platforms’ coöperation with the government threatens free speech. Then, there’s the matter of who is in charge: Elon Musk has taken over Twitter (now X), and TikTok is owned by the Chinese. Moreover, Warner told me, political polarization has made voters increasingly credulous about fake claims that reinforce their instincts—whether the subject is a stolen election or the Vance family couch.

Voters have a limited number of ways to learn about the illicit attempts of foreign powers to manipulate them. One way is for private companies—Microsoft is currently the most active—to publish research about suspicious social-media content or cybercrimes that appear to be state-sponsored. Brandt described such civilian-identified plots as “caught in the wild.” But private companies can never speak with the authority of the government, and, without subpoenas or spies, they also lack the same breadth of information. Watts, a former F.B.I. special agent, told me that the government is “the ultimate source of confirmation on attribution and actors.” A deepfake that Microsoft spots “may be the tip of the iceberg,” he continued, and U.S. intelligence officials “can understand it at a much deeper level.” Then, there is what he called “a chicken-and-egg problem” facing private companies. The government asks them to shut their platforms to malicious foreign trolls, but the companies “are waiting for the state to tell them who those accounts are.”

Criminal prosecutions are another way that covert foreign plots targeting an election can be exposed. Since the appointment of the special counsel Robert Mueller to investigate the Kremlin’s gambit in 2016, federal indictments have consistently provided the most detailed, and therefore potent, accounts of such influence operations. This past summer, news reports about a hacking of the e-mail accounts of Roger Stone, a former Trump adviser, evidently prompted prosecutors in Washington, D.C., to file an indictment against three Iranians. They were charged with dozens of hacking attacks during a five-year period, almost all of them against Americans involved in national security or foreign affairs. The U.S. government had been watching these Iranians for at least four years; the indictment cites evidence that, in each of those years, two of the operatives repeatedly visited a Tehran address linked to the crimes. On June 27, 2024, according to the indictment, the Iranians e-mailed two Biden campaign officials a stolen copy of materials that Trump had used to prepare for that night’s Presidential debate. (The Iranians presciently warned that, if Biden lost the debate, the Democrats “will have to replace” him.) There’s no evidence, however, that the recipients read the e-mails; Biden flailed in any case. A subsequent attempt to give journalists stolen vetting materials about Trump’s running mate, J. D. Vance, also found no takers. The Times reported its editors had concluded that “publication was likely to serve the interests of the attackers.”

Other legal findings, also unsealed in September, described a sweeping Russian operation that was years in the making. An affidavit by an F.B.I. investigator quoted notes from meetings held at the Kremlin by a top aide to Putin as early as April, 2022. The aide had hired three Russian contractors to conduct a covert online propaganda campaign to weaken global support for Ukraine’s attempt to repel Russia’s invasion. In 2023, one of the contractors submitted a more detailed proposal, called the Good Old U.S.A. Project, to sway the 2024 election in America. The proposal asserted that an isolationist view of the Ukraine war had become a “centerpiece” of the Presidential race; Russia must therefore “put a maximum effort to ensure that the Republican point of view (first and foremost the opinion of Trump’s supporters) wins over the U.S. public opinion.” (The names of the parties and candidates were redacted in the filing.) The proposal’s authors saw an opportunity in “the high level of polarization of American society,” which had created an “information situation” that “differs dramatically from that in all other Western countries.”

The Good Old U.S.A. Project envisaged setting up hundreds of fake online accounts, including eighteen seemingly apolitical “sleeper” groups on multiple social-media platforms across six swing states; “at the right moment,” they would “distribute bogus stories disguised as newsworthy events.” (Kremlin documents included in the filing describe Twitter as the most hospitable “mass platform,” although a partially redacted sentence suggests that the Russians liked Trump’s Truth Social even more.) To avoid detection, the Russians planned to disseminate misinformation by inserting comments or replies into authentic message threads; these comments would include links directing users to sites showcasing more elaborate propaganda. The Russians also set out to secretly promote real American influencers who supported “ending the war in Ukraine” and were “ready to get involved in the promotion of the project narratives.”

In March, two of the Russian contractors were sanctioned by the Treasury Department for their role in the operation. In July, U.S. prosecutors, after receiving a tip from another government agency, seized nearly a thousand X accounts allegedly tied to a Russian “bot farm” that used A.I. “to create fictitious social media profiles,” evidently as part of the same scheme. Finally, in September, the government shut down thirty-two Web sites that disguised Kremlin propaganda as content from news organizations such as Fox News and the Washington Post. At the same time, prosecutors charged two Russian spies with conspiring to pay ten million dollars to a group of conservative American influencers. Although the unsealed indictment redacted the names, other details indicated that the Russians worked through a Nashville startup called Tenet Media. According to the indictment, in recent months the Russians had posted nearly nine hundred video clips of their own propaganda directly to Tenet social-media feeds. Until the indictment was unsealed, American viewers had no way of knowing that the Kremlin was behind this.

But U.S. intelligence agencies definitely did, just as they plainly knew about the disguised Web sites. Details from the indictments make clear that federal prosecutors were aware of the underlying schemes for months or longer before informing voters. Of course, educating voters about foreign plots is not the primary responsibility of law enforcement, which moves at its own methodical pace. Subpoenas must be obtained to legally acquire information that other agencies might have learned through spycraft; it takes time to squeeze conspirators to testify against one another, and to lock down conclusive evidence before unsealing charges. Law-enforcement agencies may also want to delay an indictment so that they can arrest suspects before they can flee—although, in the recent election-influence cases, the three Iranians and two Russians indicted were already far out of reach.

Prosecutors also work under their own deadlines. Justice Department policy precludes the agency from taking any public actions in the sixty days before an election which might affect the outcome—including filing indictments that expose a foreign adversary’s backing of a candidate. Prosecutors appear to have kept working on the Russia indictments in secrecy as long as they could. They were unsealed on September 4th, on the eve of the sixty-day deadline. Still, Brandt told me that, whatever the timing constraints, the Justice Department can “go much farther than we can” when releasing information. “That is how you end up making public multiple internal Russian planning documents, which is something the intelligence community could never release.”

For voters, the Russia and Iran indictments also raise questions about what else the government knows. Both filings offer keyhole views of major influence operations that surely were not limited to a few inconsequential hacks and to the staff of a small Tennessee media company. Watts, of Microsoft, told me that the government is cracking down on covert Russian influence operations more aggressively than it did before the 2020 election, when there were no such indictments; prosecutors have gone after a “sizable chunk of the Russian efforts we have noted.” But he said that law enforcement had not yet taken any visible action against two other Russian online networks that Microsoft had spotted meddling in the election. The company calls those two networks Storm-1516 (which pushed the staged video falsely accusing Harris of a hit-and-run) and Storm-1679 (which pushed a viral video showing a fake New York billboard that hyped false claims about Harris).

Representative Jim Himes, of Connecticut, the ranking Democrat on the House Intelligence Committee, told me he was “quite certain” that the foreign corruption of Tenet Media was not an isolated incident: “We are going to find out there are other cases where some cutout says, ‘Hey, I’ve got five million dollars for you to promote that Fauci is a Bolshevik,’ or whatever, and the answer is ‘Yeah, give me that five million!’ ”

Hearing directly from the U.S. intelligence agencies is the third way Americans can learn about foreign efforts to manipulate our elections. This election season, the Foreign Malign Influence Center has scheduled periodic “updates” to address the torrent of questions from journalists about such plots. For the spy services, one official told me, this level of public disclosure “is like standing there naked compared to what we have done in the past.” The agencies, always zealous about protecting their sources and methods, prefer to talk as little as possible, and as vaguely as possible. The resulting updates, typically about five hundred words each, are exasperatingly abstract. Speaking as the “intelligence community,” or I.C., an update from early October noted:

A range of foreign actors continue to try to influence U.S. elections as we approach November. These activities include broad efforts aimed at undermining trust in U.S. democratic processes and exacerbating divisions in our society, while also seeking to shape voter preferences toward specific candidates. Our assessments about the activities and goals of Russia, Iran, and China are unchanged from earlier election security updates. On the presidential race, the IC continues to assess that Russia prefers the Former President and Iran prefers the Vice President; China is not seeking to influence the Presidential election.

The center also holds hour-long conference calls with journalists, but the officials on the calls limit their answers to the contents of the written updates.

The opacity of such intelligence assessments, whether to journalists or to lawmakers, inevitably opens opportunities for political spin. In 2019, intelligence officials appointed a career spy named Shelby Pierson to the new post of election-threats executive. Her job was to coördinate the analysis of foreign interference or influence operations. After Pierson briefed the bipartisan leaders of the congressional intelligence committees, people on Capitol Hill leaked that she had said the Kremlin once again preferred Trump. The President exploded in anger, tried to get Pierson fired, and attempted to stop the briefings.

She survived. But Trump then appointed two new directors of National Intelligence, both of whom downplayed the Russian threat. The first was the former ambassador Ric Grenell, who served as temporary acting director. Under Grenell’s tenure, a declassified update provided to the committees declared that the intelligence community “has not concluded” that the Kremlin was aiding either Trump or Biden, “nor have we concluded that the Russians will definitely choose to try to do so in 2020.”

John Ratcliffe, a Republican congressman and a former prosecutor, took over as director in May, 2020. He played up supposed intelligence about a major plot by China instead of Russia. Shortly before the election that fall, Ratcliffe was asked in an interview on Fox News whether China opposed Trump. Ratcliffe replied that he could not “get into a whole lot of details” in an unclassified setting, but did say that China was “using a massive and sophisticated influence campaign that dwarfs anything that any other country is doing.”

Democrats complained that the Trump appointees were twisting the conclusions of the career analysts, but the classified nature of the reports left no way to settle the dispute. Then, on January 6, 2021, the spy agencies’ “analytic ombudsman” released a report saying that, in the final year of the Trump Administration, intelligence about foreign efforts to influence the election had been “delayed, distorted, or obstructed” for “political reasons,” and that career analysts viewed some of the public statements issued under Grenell and Ratcliffe as a “gross misrepresentation” of the agencies’ assessments of the Russian and Chinese operations. (Grenell told me that the ombudsman’s report had relied on liberal partisans inside the intelligence agencies; Ratcliffe defended his statements about China as a dissenting view based on his own analysis.) Two months after Biden took office, a declassified version of the agencies’ post-election assessment stated that several arms of the Russian government had, in fact, carried out influence operations “supporting former President Trump” and that the Russians had also been spreading misinformation denigrating Biden for at least six years. A headline in the assessment declared, “China Did Not Attempt to Influence Presidential Election Outcome.”

Grenell has since become an informal adviser to Trump’s 2024 campaign, and he argued to me recently that Biden Administration appointees were now slanting intelligence about foreign influence operations to benefit the Democrats. “You’re surprised?” Grenell asked me incredulously. He added, “Putin says he would prefer Joe Biden, or now Kamala Harris, because they are more predictable! Why would you dismiss that?” (Prosecutors unsealed their detailed Russia indictment a few weeks after I interviewed Grenell, and news reports that Trump had stayed in touch with Putin after leaving the White House emerged after that.) If Trump wins, Grenell, Ratcliffe, and Turner, the chairman of the House Intelligence Committee, are all prime candidates for senior roles in the new Administration.

Brandt, the Foreign Malign Influence Center’s director, told me that the intelligence agencies now adhere to a formal protocol designed to keep politics out of the process—thereby insulating Presidents from the anxieties that stifled Obama, and from the accusations of bias that have hung over Trump and Biden. The rules, which are little known to the public and are all but ignored by the political class, were formulated in 2019, initially under the tenure of Dan Coats, Trump’s first director of National Intelligence. Coats, a former Republican senator, remains widely respected by lawmakers of both parties for his handling of that role. Biden signed off on the protocol with only slight modifications.

The process hinges on an “experts’ group” of a dozen career intelligence analysts or other civil servants from across the relevant agencies. Brandt—who was tapped for her job by the current director of National Intelligence, Avril Haines, a Biden appointee and an Obama Administration alumna—is excluded. Under the policy, the committee evaluates any intelligence of an imminent foreign-influence threat according to five criteria. Two of the criteria address the quality of the intelligence: Is it credible and specific? Three address the nature of the threat: Is it foreign in origin, underhanded or covert in nature, and severe in its potential impact? If the experts deem all five criteria met, the group can recommend a public notification.

One catch, however, is that the public does not know who sits on the experts’ group—all its members are anonymous, as is its chair. Since the Foreign Malign Influence Center was inaugurated, the intelligence authorities have withheld even the name of the election-threats executive, making Brandt (or Haines) the face of any public notifications.

Another catch is that, before a warning from the experts can be shared with the public, their recommendation must be reviewed by a “leaders’ group” composed entirely of political appointees. The group essentially duplicates the National Security Council: the director of National Intelligence convenes the Secretaries of State, Defense, Treasury, and Homeland Security, along with the Attorney General and the directors of the F.B.I., the C.I.A., the N.S.A., and the Cybersecurity and Infrastructure Security Agency. An emergency notification becomes public only with the approval of these leaders. (A loophole: the notification protocol does not necessarily restrict a director of National Intelligence from making statements or giving interviews about election threats, as Ratcliffe did.)

The day I visited the center, Brandt, in an attempt to dispel doubts about partisanship, took the exceptional step of introducing me to the chair of the experts’ group—a stern veteran of the intelligence agencies who looked at least a decade older than Brandt, and whom I agreed not to name. She told me, “As a career civil servant, I try not to have a public persona.” During her two years leading the experts’ group, she said, nobody had ever discussed potential political repercussions: “Never in any of the meetings has it even come up—what will this mean for a political party, or what will it mean for an Administration?”

The main threshold for a public notification about a piece of intelligence, the experts’ chair said, is “Could it undermine the credibility of an election or potentially change its outcome?” Brandt, speaking as a former think-tank scholar, noted that political scientists still have no accepted way to gauge the impact of an online propaganda campaign. But the chair struck a firmer tone, saying, “If we think the activity might undermine the credibility or affect the outcome, we are going to weigh very seriously a public notification.”

The experts’ chair insisted that in this cycle the intelligence agencies had not withheld information “that met all five of the criteria”—and did not risk exposing sources and methods. Nor had the leaders’ group ever overruled a recommendation by the career experts. And if they did? It would be the job of the chair of the experts’ group to stand up or speak out, she told me: “That is why we pick a career civil servant who is retirement-eligible.” In other words, she can resign in protest.

Brandt said that, if a private player like Microsoft calls out a foreign influence operation, that can alleviate the need for a government notification. In other cases, she said, law-enforcement agencies tell their intelligence counterparts, “We’ve got this one.” And if a foreign operation aims at only an individual or a campaign—as is often the case—officials from the intelligence agencies may notify the target privately.

As a result, since 2019, the experts have proposed only three public notifications. All were carried out, and all were about Iran. The first occurred on October 21, 2020, when Ratcliffe, the director of National Intelligence, publicly announced that Iran was secretly behind a wave of e-mails, putatively sent by the Proud Boys, telling Democrats that if they didn’t vote for Trump “we will come after you.” At a press conference, Ratcliffe declared that the e-mails were an Iranian plot “to incite social unrest.”

Yet Ratcliffe went on to say that Iran also sought to “damage President Trump”—a conclusion that intelligence officials told me was Ratcliffe’s own inference. Playing up the Iranian threat, he added that “we have not seen the same actions from Russia.” Democrats, fearing that Trump might gain from the impression that Iran backed Biden, spun the revelation in another direction: in a television interview, Senator Chuck Schumer, the head of the Democratic caucus, insisted that his intelligence briefing had characterized the Iranian operation as a ploy “to undermine confidence in elections, and not aimed at any particular figure.”

The second and third expert-group notifications, which took place within a few weeks of each other this year, did not forestall controversy, either. A notification issued on August 19th confirmed earlier news reports that Iran had hacked Roger Stone’s e-mails in an attempt “to compromise former President Trump’s campaign.” But the notification, unlike those reports, also brought up the Democrats. To influence the “election process,” the notification added, Iran had also sought to access “the presidential campaigns of both political parties.” A few weeks later, another notification revealed that the Iranians had sent Trump’s debate-prep materials to the Biden campaign.

Grenell, Trump’s former acting director of National Intelligence, told me that the August 19th notification’s mention of “both parties” had been a favor to Kamala Harris: the gratuitous reference to an attack on her campaign had obscured the broader fact that Iran wanted her to win. But the chair of the experts’ group defended the assessment to me, insisting that the agencies had disclosed the hacking activities as soon as they learned about them—and not in response to news reports about Stone. “We go with what we know,” she said, and argued that withholding the information about “both parties” would have been the truly partisan choice.

Both Brandt and the experts’ chair contended that the public-notification procedure was as insulated as possible from the appearance of political influence, given that the U.S. government is headed by an elected official. Nonetheless, in two out of two Presidential election cycles, the protocol has failed to allay accusations of a partisan agenda. I could see why. As I spoke with the two officials, I couldn’t shake the feeling that I was sitting across a table from people who knew much more than they were telling me about how foreign spies were trying to influence my vote or mess with our heads. My questions kept colliding with the intelligence agencies’ concern about protecting their sources and methods.

But even a little more real-time transparency would surely bolster public trust, if only by dispelling some of the mystery. Could the U.S. intelligence agencies have told the public any sooner that Iranian hackers with a history of conventional espionage were attempting to breach the Trump campaign? Did the government need to wait until almost exactly sixty days before the election to warn voters that the Kremlin was behind Tenet Media (whose YouTube videos in the past year have logged sixteen million views)? Himes, the top Democrat on the House Intelligence Committee, told me that timely information about election-influence operations too often gets bottled up by the “tension between law enforcement and intelligence gathering.” He added, “Law enforcement wants to put people in jail. Intelligence would like the criminals to keep doing what they do for twenty years, so they can identify their associates. But maybe there should be more of a tension between prosecution and the public’s right to know.”

At the end of last year, the intelligence agencies released a public version of their assessment of foreign influence operations during the 2022 midterm-election season, and it underscored how little information the government shared with Americans before voters went to the polls. The assessment described an upward trend in activity by “a diverse and growing group of foreign actors,” which the agencies attributed to “perceptions that election influence activity has been normalized” and to “the low cost but potentially high reward of such activities.” The foreign mischief that had been detected in 2022 included “payments to influencers and enlistment of public relations (PR) firms” and efforts aimed at “amplifying authentic U.S. public narratives.” Like devious music producers, foreign powers were turning up the volume of certain “authentic” American voices to maximize discord. But which influencers took what payments, and how were narratives amplified? What other governments were in that “growing group of foreign actors”? Later in the assessment, bullet points name six foreign governments—all blacked out—whose covert influence activities “did not clearly meet” an intelligence-community threshold for public disclosure. People familiar with the classified assessment told me that the redacted names were often “frenemies,” such as Middle Eastern clients with their own agendas in Washington. (Senator Warner told me, “There are countries that are our friends one day and our challengers the next.”)

Although the Foreign Malign Influence Center has said that China is staying out of the Presidential race, the center’s updates have also said that the country is attempting to sway certain down-ballot races, including “tens” of congressional races. The assessment of the 2022 election also concluded that Chinese authorities had “tacitly approved efforts to try to influence a handful of midterm races” and had “identified specific members of Congress to punish for their anti-China views.” That included “covertly denigrating a named U.S. Senator online using inauthentic accounts.” Did anyone notify the voters in the senator’s state? The intelligence officials declined to say. (The Washington Post, citing a researcher at Clemson University, recently reported that in 2022 Chinese-linked accounts had spread memes and tweets attacking Senator Marco Rubio, a prominent China hawk who was on the ballot that year. Rubio declined to respond to my questions.) Nor have the intelligence officials disclosed which other statewide races China has tried to influence in either 2022 or 2024. (The Post identified one current target as Representative Barry Moore, an Alabama Republican. A Chinese-linked account called him a “Jewish dog,” although he is not Jewish.)

Brandt told me that the elliptical bulletins are “setting the table” for the possibility that future operations by China or other nations might rise to a level meriting a public warning. She argued that, if the intelligence agencies alerted the public about every scrap of intelligence on an influence scheme, no matter how minor the threat, the constant notifications would lose their power to arouse public alarm. The din of suspicion could also weaken the credibility of the democratic process. “We would be blowing wind in our adversaries’ sails,” she said. Still, she insisted, no foreign nation got a free pass: “If you are a foreign actor trying to influence our elections, you are in our sights.”

Salvo, of the German Marshall Fund’s Alliance for Securing Democracy, said that he now worries about what will happen if the intelligence agencies successfully expose a major foreign influence operation in the final weeks of the Presidential race. “The closer we get to Election Day, the less I think that would even matter, because of the hyperpoliticized moment that we live in,” he said. “The director of National Intelligence could come out then with information about a Russian or Iranian information operation targeting Election Day, and you’ll have tens of millions of Americans who don’t believe it, because national political figures are out there challenging the Intelligence director!”

During the final weeks of the 2024 campaign, the experts’ group has been meeting three times a week to evaluate any intelligence about potential threats, and staying in contact on the weekends. In a measure of both the group’s vigilance and the over-all threat level, an intelligence official recently told journalists that the number of “nominations” the experts had proposed for a notification had “increased threefold” from the 2020 election. All but the two notifications about Iran were privately rendered, but Brandt and the experts’ chair told me that, unlike the Justice Department, the intelligence agencies have no rule against publicizing allegations about a foreign influence plot in the days before the vote. Their mandate is just the opposite. The experts’ chair said, “What we don’t want to do is get the information out after the election.”

Brandt then added, “We’ll all be a lot smarter in January.” ♦