Evening, September 26, 1960.
Here’s the scene: a vice president and a senator walked into a CBS studio in Chicago. One man wore a light gray suit that dissolved into the television backdrop, his face pale and glistening with perspiration from a recent bout of flu and a painful knee injury he had re-aggravated stepping out of his car. The other wore a dark suit, looked tanned and rested after a weekend of preparation, and had the instinct to look directly into the camera lens rather than at the journalists asking questions. Roughly sixty to seventy million Americans watched. And when it was over, something fundamental shifted in the nature of political leadership itself.
Radio listeners, according to the lore that persists to this day, thought Richard Nixon won. Television viewers overwhelmingly gave the debate to John F. Kennedy. As Kennedy’s internal pollster Lou Harris reported, the senator surged to a 48-to-43 percent lead after that first debate. Before the debate, Nixon had led by six points nationally.
Here’s the takeaway: a president’s image matters, but the dominant communications medium of any era does not simply transmit leadership, it shapes it, selects for it, and ultimately redefines what leadership means for that presidency. If we look at the arc of the last century through this lens, you will see a clear pattern.
In the 1930s and 1940s, radio was the intimate intruder, entering living rooms, kitchens, and bedrooms with a disembodied voice that seemed to speak to each listener personally. Franklin Roosevelt understood this with preternatural brilliance. His Fireside Chats were conversations. He addressed Americans as “my friends,” modulated his patrician tenor to sound warm and accessible, and explained complex policy, banking crises, wartime mobilization, as though tutoring a neighbor over coffee. Radio rewarded vocal warmth, narrative clarity, and the ability to project calm authority without visual distraction. Roosevelt’s wheelchair was invisible. His patrician jaw and aristocratic bearing were irrelevant. What mattered was the voice, and the voice conveyed mastery, reassurance, intimacy.
Winston Churchill, too, wielded radio as a weapon of civilizational survival. His wartime broadcasts transformed the English language itself into a rallying cry. Radio leadership was fundamentally about inspiration, the capacity to summon courage, articulate shared sacrifice, and project moral certainty through the spoken word alone.
Then television arrived, and everything changed. The medium that rewarded vocal intimacy gave way to one that rewarded visual charisma, composure under pressure, and the ineffable quality of telegenic authority. Kennedy embodied this transition so completely that three subsequent sitting presidents, Lyndon Johnson, who understood his own limitations before the camera, Richard Nixon, and Gerald Ford, refused to participate in televised debates for the next sixteen years. It was Ford who finally restored the tradition in 1976, and promptly demonstrated television’s unforgiving nature by declaring, to a national audience, that there was “no Soviet domination of Eastern Europe.”
Television favored style over substance and even created a new kind of substance, one in which composure, physical vitality, the ability to project empathy through a screen, and the instinct for visual storytelling became genuine leadership competencies. Kennedy’s practice of looking into the camera rather than at the questioners, making viewers feel he was speaking directly to them, was a new form of democratic communication.
By the 1990s, television had evolved from the formal broadcast stage of Kennedy and Nixon into the late-night couch and the town hall. Bill Clinton grasped this shift with the same intuitive brilliance Roosevelt had brought to radio. When Clinton appeared on The Arsenio Hall Show in June 1992, donning sunglasses and playing saxophone, he demolished the incumbent George H.W. Bush’s greatest advantage, the dignity and gravitas of the presidency, by redefining what presidential bearing could look like. Against a sitting president who was widely perceived as out of touch with ordinary Americans struggling through recession, Clinton’s ease with popular culture signaled something voters craved: a leader who lived in the same world they did.
The town hall debate format of October 1992 crystallized the contrast. When an audience member asked the candidates how the national debt had personally affected them, Bush glanced at his watch, an image that became iconic for its suggestion of disconnection. Clinton walked toward the questioner, made eye contact, and said he felt her pain. Television had evolved from the broadcasting of authority to the performance of empathy, and the leaders it selected for had evolved accordingly.
Then came the internet, and with it, a revolution in the architecture of political power that made all previous media transitions seem incremental.
Barack Obama’s 2008 campaign used the internet by reimagining the relationship between a candidate and the citizenry. By Election Day, the Obama campaign had gathered more than thirteen million email addresses and built an online platform where supporters could donate money, set up their own fundraising campaigns, organize events, make calls to battleground states, and work within volunteer groups in their communities. The campaign raised approximately $750 million total, with roughly $500 million coming through digital channels. Millions of individual donors contributed, many giving under $100 each, a small-donor base that would grow further in the 2012 reelection, when the campaign’s email program alone brought in roughly $485 million.
Read: Two wars, one nation, citizens divided (25, 2026)
This was a philosophical transformation, and not merely a fundraising innovation. The internet democratized the donor base, turning political participation from a spectator sport into a networked community of stakeholders. Obama’s digital team, led by figures like Teddy Goff and the firm Blue State Digital, pioneered techniques of micro-engagement, personalized emails, and culturally targeted outreach that created an emotional connection and a continual conversation between candidate and citizen. The internet, in short, rewarded leaders who could build movements, who could inspire not just audiences but participants. It selected for authenticity, grassroots energy, and the capacity to make millions of people feel they had personal ownership of a political outcome.
And then social media swallowed the internet whole, and the rules changed again. This time it has changed in ways that raise questions not merely about leadership style but about the survival of democratic governance itself.
Donald Trump’s ascent was inseparable from the architecture of social media platforms, particularly Twitter, which rewarded provocation, controversy, and the relentless compression of complex issues into combative slogans. Trump understood something profound about the attention economy: that in a media ecosystem driven by algorithmic amplification of engagement, conflict is content. His rallies were reality television. His tweets were cliffhangers. His feuds were plot arcs. The skills that had made him a ratings success on The Apprentice, dramatic confrontation, decisive judgment rendered in simple declarative sentences, the performance of strength, translated seamlessly into a political landscape where the currency was not policy sophistication but emotional intensity.
Social media changed both the style of leadership and the incentive structure. Platforms algorithmically rewarded content that generated strong emotional responses, outrage, fear, tribal solidarity, over content that informed, nuanced, or deliberated. And it created a political environment in which activating tribal divisions became more strategically rational than inspiring broad coalitions.
It is now harder to inspire than to inflame.
Roosevelt inspired a nation through depression and war. Kennedy inspired a generation to public service. Obama inspired a multiracial coalition to believe in democratic renewal. But each of these leaders operated in media environments that, for all their flaws, retained some structural incentive for broad appeal. Radio and network television were mass media. They reached everyone, and the leaders who thrived in them needed, at minimum, to speak to a general audience.
Social media is fragmented media. It sorts citizens into affinity groups, feeds them content calibrated to their existing beliefs, and creates information ecosystems so divergent that citizens in the same democracy may inhabit entirely different realities. In such an environment, the rational strategy for a political entrepreneur is not to bridge divides but to deepen them, not to expand a coalition but to intensify the loyalty of a tribe.
Now add artificial intelligence, and with it, a threat that makes the social media distortion look quaint.
We are now living in the deepfake era. The 2024 election cycle became the first major American contest to unfold in the era of sophisticated generative AI, and campaigns deployed AI-generated content with startling aggression. In June 2023, during the Republican presidential primary, Ron DeSantis’s campaign released a video splicing in AI-fabricated images of Donald Trump embracing Anthony Fauci. A fake robocall impersonating President Biden advised New Hampshire Democrats not to vote in the primary, a stunt the FCC later fined at $6 million. In the 2025 New York City mayoral race, Andrew Cuomo’s campaign released an AI video depicting opponent Zohran Mamdani alongside a montage of criminality, including racially charged imagery, that was pulled within minutes of posting but not before being reshared thousands of times. The incident became a case study in how AI-generated attack content, even when hastily retracted, can leave lasting impressions.
The volume of deepfakes has grown explosively: cybersecurity researchers estimate an increase from roughly 500,000 online deepfakes in 2023 to about 8 million in 2025, with annual growth nearing 900%. The qualitative shift is as alarming as the quantitative one. Voice cloning has crossed what researchers call the “indistinguishable threshold.” A few seconds of audio now suffice to generate a convincing clone, complete with natural intonation, rhythm, emphasis, emotion, pauses, and breathing. Human detection rates for high-quality video deepfakes stand at just 24.5 percent, and a 2025 iProov study found that only 0.1 percent of participants correctly identified all fake and real media shown to them.
A study of the “liar’s dividend” found that deepfakes are approximately as credible as other forms of misinformation communicated through text or audio. But researchers uncovered something more insidious: subjects frequently confused authentic videos for deepfakes when the content depicted an elite from their own political party in a scandal. This is the “liar’s dividend,” a concept articulated by scholars Danielle Citron and Bobby Chesney in which the mere existence of deepfake technology allows anyone caught in genuine wrongdoing to dismiss authentic evidence as fabricated. The erosion of shared evidentiary reality may ultimately prove more damaging to democracy than any individual fake video.
By mid-2025, 47 states had enacted laws targeting AI-generated synthetic media. The federal TAKE IT DOWN Act, signed into law by President Trump on May 19, 2025, following near-unanimous congressional support, 409 to 2 in the House and unanimous in the Senate, provides the first nationwide framework addressing intimate deepfakes. But as First Amendment scholars have observed, the constitutional terrain is treacherous. A federal judge struck down California’s pioneering AB 2839 as unconstitutional, finding that its broad prohibition on AI-generated election content conflicted with First Amendment protections for political speech. The lesson is sobering: legislation is racing to catch a technology that the courts keep reminding us resembles speech.
Read: The Long Gaze: From Paris to NY to Bombay (19, 2026)
Back to debates, will they matter now in this deepfake era?
The presidential debate format was born in a media environment where tens of millions of citizens shared a single experience simultaneously, where the camera was an impartial witness, and where a candidate’s composure, knowledge, and character could be assessed in real time by the entire electorate. That world is vanishing. When voters receive their political information through algorithmically curated feeds, when AI can generate convincing video of a candidate saying anything, when a well-timed deepfake released hours before an election may be impossible to debunk before polls close, the traditional debate becomes a relic, a civic ritual whose power depends on a shared information environment that no longer exists.
Or, precisely because the information environment is fragmenting, the live debate, unedited, unfiltered, happening in real time, may become one of the last arenas where citizens can assess leadership directly. This assumes that enough citizens will watch. That what they see will not be mediated through so many layers of partisan interpretation that the shared experience becomes an illusion.
Meanwhile, America’s adversaries are watching and they have recognized that the same technologies reshaping our domestic politics can be weaponized against us from abroad. Foreign adversaries can now create synthetic identities, complete with AI-generated profile photos and deepfake voices, at industrial scale, inject fabricated content into American political discourse through fake grassroots groups and spoofed news outlets, and exploit the same algorithmic amplification that rewards domestic political polarization.
Meta has been combating the “Doppelganger” campaign, which created elaborate websites spoofing news organizations including The Washington Post and Fox News, then disseminated links through networks of fake social media accounts. The volume of posts, articles, and websites that Russian-linked operations produce is being boosted by artificial intelligence, another factor that sets recent election cycles apart from earlier ones. Covert influence campaigns based in Russia, as well as China, Iran, and Israel, have begun using AI in their attempts to manipulate public opinion. In 2023, Meta dismantled what it identified as the largest Chinese covert influence operation in the company’s history, linking it to Chinese law enforcement and tracing a sprawling network of fake accounts across more than 50 websites, from Facebook and Instagram to YouTube, TikTok, Reddit, and dozens of smaller platforms. North Korea, too, has developed sophisticated cyber capabilities that increasingly blur the line between espionage, financial crime, and information warfare. The FBI has documented North Korean state-sponsored actors impersonating journalists and academics to infiltrate think tanks and policy circles, using social engineering techniques that exploit personal details gleaned from social media to craft highly personalized deception.
And what about money? Crypto? Dark super PACs, and anonymous political influence?
The 2024 federal election cycle was the most secretive since the Supreme Court’s Citizens United decision in 2010. Dark money groups, nonprofits, and shell companies that spend on elections without revealing their donors plowed more than $1.9 billion into the cycle, a dramatic increase from the prior record of $1 billion in 2020. Total outside spending on the 2024 elections hit a record $4.5 billion, with more than half of that spending coming from groups that do not fully disclose their funding sources. Independent political committees received over $1 billion from shell companies and dark money nonprofits not required to disclose their donors, a figure that more than doubled from the 2020 cycle’s $653 million.
Now cryptocurrency has emerged as a political force of the first order. Fairshake, the crypto industry-backed super PAC, disclosed holding $193 million in cash on hand as of January 2026 ahead of this year’s midterm elections, fueled by $25 million from Ripple Labs and $24 million from Andreessen Horowitz. By late June 2026, that spending had accelerated further: a Public Citizen analysis found crypto companies had already poured $189 million into the 2026 midterms, making the industry the single largest corporate political spender of the cycle, while independent trackers put total crypto-aligned super PAC spending, including Fairshake and its affiliates, above $270 million. Fairshake’s largest donors remain Ripple Labs and Coinbase. Coinbase directed roughly $86 million to Fairshake during the 2024 cycle alone, and by mid-2026 had contributed a further $33 million toward the 2026 midterms. The most prominent dark money group tied to the crypto industry is the Cedar Innovation Foundation, which does not disclose its donors and shares a spokesperson with Fairshake.
If you have read this far, you can now see the structural vulnerability.
Cryptocurrency’s design principles, decentralization, pseudonymity, borderless transactions, make it an ideal vehicle for those seeking to influence elections while concealing their identities. When combined with the existing dark money infrastructure of shell companies funneling anonymous contributions through 501(c)(4) nonprofits to super PACs, cryptocurrency could further obscure the already murky trail between donor and candidate, including, potentially, foreign donors prohibited by law from participating in American elections.
So where does this leave us?
A century of leadership evolution, driven by successive communications revolutions, has brought us to a moment of profound democratic vulnerability.
Radio selected for leaders who could inspire through voice. Television selected for leaders who could project authority and empathy through image. The early internet selected for leaders who could mobilize communities and democratize participation. Social media selected for leaders who could activate tribal loyalties and generate emotional engagement. And now artificial intelligence threatens to select for leaders who can most effectively deploy fabrication, or, perhaps worse, to make the distinction between authentic and fabricated leadership impossible for citizens to discern.
See the pattern? Each successive medium has shifted the balance further from substance toward performance, from deliberation toward activation, from the broad public square toward the fragmented tribal enclave. And at each transition, the same question recurs: can democratic self-governance survive a communications environment that rewards manipulation over truth?
I am not a pessimist by nature. I believe in the resilience of democratic institutions and the ultimate wisdom of an informed citizenry. But I am also a realist. Gartner has projected that by 2026, 30 percent of enterprises will no longer consider standalone identity verification reliable in isolation, a warning that speaks not only to the corporate world but to the electorate at large. When adversarial nations can inject fabricated content into our political discourse at scale, when anonymous money flows through crypto-enabled dark channels to amplify that content, when the very concept of evidentiary reality is being eroded by technologies that make seeing no longer equivalent to believing, the foundations of self-governance are under assault in ways that no previous generation of Americans has confronted.
The antidote, if there is one, lies not in nostalgia for the Fireside Chat or the town hall debate, but in building new institutions of epistemic trust, verification systems, media literacy initiatives, transparency requirements for AI-generated content, campaign finance reforms that trace money to its human source, that are adequate to the technological moment. It lies in citizens who refuse to surrender their capacity for critical judgment, and in leaders who understand that the ultimate measure of leadership is not the medium through which it is communicated, but the character, the competence, and the commitment to truth that it embodies.
Kennedy, for all the mythology of that first debate, won the presidency by the narrowest of margins, 49.7 percent to 49.5 percent in the popular vote. Television did not hand him the election. What it did was create a new arena in which his particular gifts could be displayed, and in which voters could make a judgment, imperfect, influenced by image, but ultimately their own.
Will we allow the new technologies of fabrication to foreclose the possibility of democratic choice itself?
To be continued…