Moscow has dramatically tightened the guardrails on artificial intelligence in political life, as the upper house of Russia’s parliament approved legislation prohibiting the use of AI-generated images and voices of individuals in election campaigning without explicit written consent. The Federation Council’s vote on 29 April, which followed swift passage through the State Duma, also forbids the use of AI-created likenesses of fictional and deceased persons – a measure that effectively eliminates the spectre of synthetic ‘ghost’ candidates. Viewed from Moscow, the law is framed as a defence of electoral integrity against disinformation, yet it also grants incumbents and state-aligned entities precise control over which faces and voices populate the public square.

The new regulations are layered with conditions that betray Russian political realities. Any candidate or party may deploy AI-generated versions of their own nominee, but only if the individual is a Russian citizen over the age of eighteen and has provided written authorisation. More pointedly, if a person depicted or heard in campaign material is deemed “affiliated with a foreign agent,” that label must be prominently displayed. This requirement, analysts in Brussels note, extends the Kremlin’s decade-long effort to stigmatise voices it considers outside influence, now blending it with the misinformation anxieties that preoccupy Western democracies.

The Russian move arrives amid a wider global scramble to set legal boundaries for synthetic media. On the other side of the Atlantic, the singer Taylor Swift has pursued a parallel but distinct strategy: registering her own voice and image with the United States Patent and Trademark Office. The filings, lodged on 24 April, aim to protect against AI-generated deepfakes and false political endorsements that have proliferated online. Swift’s initiative exposes the yawning gaps in US federal law, which offers no comprehensive right of personality, leaving celebrities to cobble together protections from trademark, copyright and state-level publicity laws. In Washington, the episode has reignited discussions over the proposed No Fakes Act, though legislative momentum remains sluggish.

The contrast between the Russian and American approaches is instructive. Moscow has chosen to hard-code restrictions directly into electoral law, creating a state-administered gatekeeping mechanism. Washington, true to its legal culture, relies on individual action and the courts, leaving a vacuum that campaign strategists and malicious actors can exploit. Viewed from London, both paths reflect a reactive posture: governments and public figures are only now confronting a technology that has already outpaced existing norms. With pivotal elections looming in the European Union, the United Kingdom and the United States, the race to define what is real in political communication has never been more urgent.

Looking forward, the fragmented nature of these responses points to a turbulent cycle ahead. Russia’s law, for all its restrictions, places the power of definition squarely in the hands of the state – a model unlikely to be replicated in open societies but certain to be studied by authoritarian peers. Swift’s defensive registrations, meanwhile, may spur a wave of celebrity filings that further monetise personal identity rather than truly contain deepfakes. The fundamental tension, as a seasoned observer in Geneva might conclude, lies between treating synthetic media as a tool to be disclosed and regulated, or as a weapon to be prohibited. The months before this year’s global election bonanza will test which philosophy – if either – can preserve democratic trust.