Lawfare Daily: Rethinking Deepfake Response with Gavin Wilde

Podcast: The Lawfare Podcast
Guests: Justin Sherman (Host), Gavin Wilde (Carnegie Endowment for International Peace)
Date: September 26, 2025

Episode Overview

This episode of The Lawfare Podcast features a deep-dive conversation between host Justin Sherman and Gavin Wilde, a nonresident fellow at the Carnegie Endowment for International Peace and author of the recent paper "Pyrite or Panic: Deepfakes, Knowledge, and the Institutional Backstop." The discussion explores how fears around deepfakes compare to historical anxieties over new technologies, the reality versus perception of deepfake threats—especially in the context of elections and disinformation—and what a more effective, less tech-obsessed response might look like. Wilde argues for shifting the focus toward reinforcing social and institutional backstops for knowledge, rather than becoming overly fixated on technological solutions.

Key Discussion Points & Insights

1. Deepfakes in Historical Context

• Wilde emphasizes that anxieties about new media technologies (audio, visual) are not unique to the deepfake era; historical examples show similar fears were largely overblown.

  "The line between evidence and expression, between depiction and detection... has always been very blurry."
  — Gavin Wilde [00:52]

• The "Pyrite" (fool's gold) analogy: Deepfakes are like pyrite—physically similar to gold but not the same. Historically, the discovery of pyrite increased the value placed on reliable validation methods for gold rather than destabilizing the gold market itself. Similarly, knowledge is "backstopped" by institutional and social processes more than by technology alone.
  — Gavin Wilde [04:48]

2. Evolution of Audiovisual Fakery

• Early photo and audio media were routinely manipulated, and only over time did society come to treat recordings as authentic evidence.

• Host and guest discuss how magic shows and staged recordings were accepted forms of entertainment, not considered existential threats to truth.

  "Fakery itself is not new... the line between capturing and manipulating an event has always been blurry."
  — Gavin Wilde [13:14]

• Major doomsday predictions about new media (photography killing painting, radio supplanting records, movies destroying textbooks) generally failed to materialize.

3. Deepfakes and the Current Information Environment

• The 2015–2024 surge in deepfake technology coincided with societal anxieties about disinformation, institutional mistrust, and political polarization.

  "By the time you have the first papers... about the potential harms from deepfakes in 2015, we already had a very fractured media landscape... we were already well primed... to think first and foremost about how they might be weaponized."
  — Gavin Wilde [18:15]

• Wilde argues that deepfakes have, in practice, been used more for satire and artistic expression than for effective mass disinformation.

  "The deep fakes we saw... the fakery was the point." — Quoting a YouTube executive [24:21]

• 2024, a year of mass elections worldwide, was anticipated as a test case for deepfake influence—but the reality was anticlimactic.

  • Most deepfakes were easily spotted or intentionally absurd.
  • No significant new demand for politically weaponized deepfakes emerged.
  • Notre Dame research confirmed most "fake" content online was effectively memes or satire, not deceptive propaganda.

4. Misplaced Policy Attention and The Liar’s Dividend

• Much academic and policy focus is on the speculative "collapse of knowledge" due to deepfakes, but this may distract from actual, ongoing harm such as nonconsensual intimate imagery and voice cloning scams.

  "Despite now almost a decade of theorizing about... societal epistemological impact of deepfakes... the nonconsensual imagery of women and children... voice cloning... these are at present still pretty much regulated with a patchwork of state statutes, common law principles, but we don't have much federal or national level attention."
  — Gavin Wilde [37:07]

• The legal and policy landscape is fragmented—with strong laws against child sexual abuse material (CSAM) in 45 states but little unified federal action on broader deepfake harms.

5. Audiences and the Myth of Deepfake Persuasion

• There’s disproportionate focus on whether deepfakes are technically convincing, not on the audience’s context and willingness to be “duped.”

  "We don't survey folks coming out of an action movie... Did it convince you that it was real life? No, because the context... are not the same."
  — Gavin Wilde [31:16]

• Wilde likens engagement with deepfakes to professional wrestling (kayfabe)—most people knowingly participate, with the entertainment value outweighing concerns about authenticity.

• The most effective way to safeguard knowledge is through reinforcing societal and institutional “mining and assaying”—library science, archiving, credentialing, public journalism—rather than relying solely on technological detection or watermarks.

  "If overreliance on audiovisual media as evidence brought us to this place of alarm... doubling down on that bet... is probably misguided."
  — Gavin Wilde [34:17]

Notable Quotes & Memorable Moments

• On the social construct of knowledge and authenticity

  "...the worth of pure gold is ultimately a social construct... based not only on its relative scarcity in the ground, but on the customs and the disciplines involved in mining gold and assaying gold."
  — Gavin Wilde [04:48]

• On human psychology and relevance

  "Humans pretty much ignore most media content that isn't immediately and deeply relevant to them."
  — Gavin Wilde [24:21]

• On the real threats of deepfakes

  "The actual harms... [are] nonconsensual imagery, voice cloning scams... [these] need the most attention but don't seem to be getting it, particularly at the federal level."
  — Gavin Wilde [37:07]

• On the audience’s role in meaning

  "We're all wise to the act and we choose to be here anyway... actual persuasiveness hinges [on] context and how relevant it is to a human being's immediate circumstances."
  — Gavin Wilde [31:16]

Important Segment Timestamps

• Blurry Boundaries of Evidence: [00:52]
• Pyrite/Fool’s Gold Analogy: [04:48]
• Media History Lessons: [09:42]
• Historical Fakery in Audiovisual Media: [13:14]
• Disinformation & Deepfake Panic circa 2015-2024: [18:15]
• Election Year 2024 — The Dog That Didn't Bite: [24:21]
• Platform Takeaways & Content Moderation: [27:57]
• Audience Context vs. Believability: [31:16]
• Policy Mischaracterization and Real Harms: [37:07]

Conclusion & Recommendations

• Recalibrate focus: Move away from panic over “collapsing knowledge” and the technological cat-and-mouse of detection, in favor of reinforcing institutional and social frameworks for knowledge validation.
• Policy action: Address real, ongoing harms (nonconsensual deepfakes, voice cloning) with coherent legal frameworks, particularly at the federal/national level.
• Research directions: Study not just how realistic deepfakes appear, but how audiences interact with and interpret them—moving beyond the question: "Can you spot the fake?"

Summary prepared for listeners seeking actionable insights and context on deepfakes, media trust, and the path forward.