Technology & AI
Should AI-Generated Political Deepfakes Be Criminalised?
LNAT Section B · Model essay
The essay prompt
Should it be a crime to create or spread AI-generated deepfake videos and audio of politicians during election campaigns? Set out the arguments on each side and reach a reasoned conclusion.
The stance
Yes, but only through a narrowly tailored offence. A political deepfake counterfeits a person's identity and audiovisual 'evidence', so when it is deployed to deceive voters about a candidate or the voting process it strikes at electoral integrity in a way ordinary spin does not. Counterspeech and civil law cannot outrun a clip that travels in hours. The defensible position is therefore criminalisation that is intent-based, confined to the regulated campaign period, and carved out for satire, journalism and art; a blanket ban on the technology itself would be a different and indefensible thing.
The case for
A deepfake counterfeits perception itself, so it bypasses the reasoning counterspeech relies on.
An ordinary lie still has to persuade; a fabricated video of a candidate hijacks the viewer's senses before any argument begins. The 2024 New Hampshire robocall, which cloned President Biden's voice to tell Democrats not to vote in the primary, shows the mechanism: it did not argue, it impersonated. When fabrication operates below the level of deliberation, the marketplace-of-ideas defence, that more speech cures bad speech, runs out of time.
The case against
A broadly drafted offence would chill satire, parody and legitimate comment.
Political mockery, impressions and obviously synthetic comedy are core protected expression. A vague law against 'deceptive synthetic media' risks sweeping in 'deep-Trump' comedy or campaign attack ads that voters understand as caricature. In United States v Alvarez (2012) the US Supreme Court struck down a law criminalising lies about military honours, holding that falsity alone does not strip speech of protection. Over-broad criminalisation would punish the lie rather than the harm.
Elite unlocks every LNAT Section B model essay in full — the worked argument, steel-manned counters, verified authorities and the founder's own plan behind each one.
Sign up and unlockThe founder's own plan for this essay
The exact plan this essay was built from — the founder's working notes before the model answer was written.
Elite unlocks every LNAT Section B model essay in full — the worked argument, steel-manned counters, verified authorities and the founder's own plan behind each one.
Sign up and unlockGenerate a model essay from this plan
Turn this plan into a finished essay, instantly.
Elite turns the founder's plan into a complete, 500–600 word Section B model essay written to the structure above — a clear stance, distinct points, a steel-manned counter and a tight finish. Generate, read, regenerate.
Sign up and unlock