Episode Summary
Executive Summary: Kara Swisher interviews disinformation expert Renee DiResta about Stanford Internet Observatory’s effective dismantling under political pressure, her book on how influencers, algorithms, and crowds create “bespoke realities,” and the broader collapse of trust online. They argue propaganda is now decentralized, participatory, and driven by incentives, while platforms’ retreat from moderation and opacity make election manipulation, harassment, and real-world violence more likely.
Main Topics: Stanford Internet Observatory and political pressure (Priority: 5/5): DiResta explains SIO’s work on online abuse, election rumors, vaccines, and state-linked influence campaigns, and describes how congressional inquiries, lawsuits, and harassment helped hollow out the group and chill research. Propaganda as decentralized, participatory influence (Priority: 5/5): The conversation reframes misinformation/disinformation as propaganda shaped by influencers, algorithms, and crowds, arguing that users actively amplify content and help create reality rather than passively consume it. Platform incentives and moderation retreat (Priority: 5/5): Swisher and DiResta discuss how engagement-driven business models reward outrage, how moderation became equated with censorship, and how platforms have become more opaque and less willing to intervene. Lawfare, forum shopping, and intimidation (Priority: 4/5): DiResta says lawsuits, subpoenas, and forum shopping are being used as tools to exhaust researchers, suppress speech, and create a chilling effect even when claims are weak or pending. Elon Musk, X, and the spread of harmful narratives (Priority: 4/5): They criticize Musk’s hands-off approach and his role in amplifying harmful claims, especially in the UK, where falsehoods spread on X helped fuel violence after the Southport killings. 2024 election risks and societal cohesion (Priority: 5/5): DiResta warns that the biggest danger is not foreign trolls alone but authentic domestic actors, platform dynamics, and post-election narratives that could trigger scattered violence and threats against election workers.
Key Arguments: SIO’s work was standard academic research into adversarial abuse, child safety, election rumors, vaccines, and state influence operations—not a censorship operation. The campaign against SIO worked by turning ordinary research coordination into alleged “collusion,” then amplifying that framing through Congress, media, and lawsuits. Propaganda today is less top-down than in the mass-media era; influencers and crowds now co-produce narratives with algorithms amplifying the most engaging content. “Misinformation” is too narrow because online claims often begin as rumors or contested narratives; propaganda better captures the use of uncertainty and innuendo for political ends. Users are not passive online; virality is active because people choose to click, share, and trend content, making crowd behavior part of the problem. Platforms benefit financially from outrage and engagement, but when moderated content creates political backlash, they are incentivized to retreat from responsibility. Government and platform conversations should be transparent, including disclosure of takedown requests, but investigations are often aimed more at intimidation than policy solutions. The main election threat is domestic polarization and authentic American actors, not just foreign influence operations; this makes the problem harder because it becomes a speech and culture issue, not only a security issue. Platforms should pre-commit to rules for handling candidate incitement, threats, and election-worker harassment before crises happen. Real-world violence can be sparked online even without direct instruction, as seen in the UK riots following false claims spread on X.
Data Points: SIO start year: 2019 - DiResta joined Stanford Internet Observatory as technical research manager in 2019. June contract decision: This past June - She learned her Stanford contract would not be renewed that month. Election inquiries: 2 separate congressional inquiries - She says SIO faced two congressional inquiries led by Republican pressure campaigns. Congressional subpoenas: 91 subpoenas - She says Jim Jordan sent out 91 subpoenas in related investigations. Viral tweet allegations: 22 million tweets - Conservative groups alleged SIO colluded with Twitter to suppress this number of tweets. UK X amplification: 30,000+ mentions - A false name tied to the Southport case received over 30,000 mentions on X in under 24 hours. UK X amplification accounts: 18,000+ unique accounts - Those mentions came from more than 18,000 unique accounts. 2024 reference: 2024 - DiResta says the U.S. is facing another extremely online presidential election in 2024. SIO work window: 2020 and 2022 elections - She discusses election integrity work for both cycles, including rapid-response rumor tracking.
Pivotal Quotes: "“What you've essentially done, you destroy it.”" — Renee DiResta: On how attacks on researchers and election networks break the relationships needed to counter false narratives. "“The most important thing to be thinking about now is independent of actor.”" — Renee DiResta: On focusing less on foreign trolls and more on the broader societal mechanisms that spread lies and division. "“It’s an architecture problem.”" — Kara Swisher: On how platform design and incentives shape whether online systems amplify crazy, calm, or harmful behavior.
Implications: The episode argues that democratic resilience now depends on protecting research networks, increasing platform transparency, and designing systems that reward accuracy over outrage before election crises erupt.