What Is AI Recasting?

AI recasting is the process of replacing an on-screen performer’s face, body or voice in existing footage with a different performer, using generative AI techniques.^[1] In advertising this is often used to localise a TV commercial for new markets or to refresh talent without a full reshoot, sitting alongside face-swap localisation and AI casting in modern production workflows. The practice raises material questions about consent, usage rights and audience expectations that brands must manage carefully.^[2]

Definition and uses in advertising

In commercial production, AI recasting refers to algorithmically replacing an identifiable performer in finished footage, while keeping the original motion, framing and edit intact.^[1] It can involve facial replacement, full-body re-rendering and synthetic voice dubbing. Brands use it to update talent, extend campaigns into new markets, or feature multiple local ambassadors without re-shooting each execution.^[3] The approach is particularly common in multi-language TVC localisation, where the same master film is adapted with different faces and voices for regional audiences.

Recent high-profile campaigns have demonstrated these techniques at scale. Cadbury India’s ‘Shah Rukh Khan-My-Ad’ initiative generated thousands of localised video spots where the actor’s likeness delivered personalised shop names for small retailers, using AI to re-synthesise his face and voice from base footage.^[3] Coca-Cola and Mondelez have also explored AI-driven personalisation and localisation, using generative tools to adapt creative assets for multiple markets while retaining core brand elements.^[3][4] These examples illustrate how AI recasting is moving from experimental to planned, line-item production activity.

How AI recasting works technically

Most AI recasting workflows combine facial synthesis, neural rendering and speech technologies. Modern face- and body-swapping tools use deep generative models such as diffusion or GAN-based architectures to map the target performer’s facial geometry and appearance onto the original actor frame by frame, while maintaining lighting, perspective and expression dynamics.^[1][5] For full-body recasting, systems may reconstruct a 3D representation of the replacement performer and re-render them in place of the original, guided by the source actor’s motion.

Voice recasting typically uses neural text-to-speech or voice cloning models trained on samples of the replacement performer’s speech, producing dialogue that can match local language, tone and timing.^[5] The output is then conformed to the original lip movements using automated speech-to-lip alignment, or, in more advanced systems, by regenerating mouth shapes directly in the video.^[1][5] While consumer tools have made basic recasting accessible, professional advertising workflows usually involve human supervision for quality control, brand safety and cultural checks.

Rights, consent and regulation

AI recasting sits at the intersection of performers’ rights, advertising regulation and privacy law. In the UK, the Advertising Standards Authority has made clear that any materially altered imagery must not mislead consumers, for example by implying an endorsement or performance that did not occur.^[2] Where synthetic media could materially affect how an ad is understood, marketers are expected to ensure it is presented in a way that does not deceive.^[2] Similar principles apply in many markets under unfair commercial practice and consumer protection rules.

From a labour and contractual perspective, performers’ unions such as Equity argue that any use of an actor’s image or voice for AI training, cloning or recasting requires explicit, informed consent and appropriate remuneration.^[6] Equity’s AI toolkit advises performers to scrutinise contracts for broad rights grants, negotiate specific limits on synthetic reuse, and ensure time-bound, territory-specific licences.^[6] Brands and agencies therefore need clear clauses covering digital doubles, recasting and localisation, and should document consent from all original and replacement performers before deploying AI recast assets in market.

Practical considerations for brands

For marketers, AI recasting is essentially a post-production localisation and versioning tool. Potential benefits include reduced reshoot costs, faster turnaround for market-specific edits, and the ability to feature high-profile ambassadors in personalised or long-tail executions that would be impractical to film individually.^[3][4] However, these efficiencies must be balanced against legal review, union rules, and talent-relations considerations, particularly where celebrity likenesses are used at scale.

Governance and disclosure policies are increasingly important. Brands should maintain an internal register of campaigns that use AI recasting, document technical vendors and model inputs, and agree clear sign-off processes across legal, production and brand teams.^[2][6] Testing with consumers can help identify any risk of perceived deception or uncanny quality issues before full deployment.^[4] Used transparently, and with robust consent and rights management, AI recasting can sit alongside traditional production rather than replacing it outright.