A deepfake election

✨ Watch on YouTube or listen on Apple Podcasts or Spotify.

Four days before the 1924 general election, a letter appeared in the British press. It purported to be from a Soviet official, inciting revolution. It was fake. The government fell anyway. A century later, you no longer need a forged document, a printing press or even a budget. You need a laptop and an afternoon.

That is the opening premise of Frances Lasok’s work on AI deepfakes and elections — and it is a more sobering starting point than it might appear, because the Zinoviev letter was, by modern standards, laughably crude. What is coming is not crude at all.

Frances is a writer and policy thinker whose recent report on deepfakes and election law is among the clearest-eyed assessments of a problem that most legislators are still pretending is theoretical. Her argument is not that something new has entered the world — unscrupulous actors have always tried to mislead voters — but that the cost of doing so has collapsed to near zero.

Much of UK election law dates from the nineteenth century. Its principles — that campaign material should identify its source, that false statements about candidates are prohibited, that those who campaign on a candidate's behalf share legal responsibility — are, she argues, sound. The problem is that those principles were written for a world of leaflets and loudspeakers, and have been only fitfully extended to the online environment. It took until 2022 for a Facebook post to legally require the same imprint — the declaration of who is behind it — as a paper leaflet pushed through a door. Deepfakes represent the next frontier, and the gap is already being exploited.

In this episode, you’ll learn:

• Why the law urgently needs to catch up with deepfake technology — and why the Victorian principles underlying UK election law are actually sound, but need to be applied to a digital environment they were never designed to cover.

• Why we are in a dangerous transitional window right now — as the technology to create convincing fakes has outpaced public scepticism, creating conditions for electoral manipulation before audiences learn to treat video with the same healthy distrust they apply to unverified text.

• Why a right to likeness may be the next major legal frontier — and how the argument mirrors the historical case for copyright: a concept that became necessary only when technology made mass replication of someone else's creation suddenly, trivially possible.

Watch the episode on YouTube below, or listen via your preferred podcast app.