This work breaks new ground for Dove as a brand, for the issue it spotlights, and for the technology it utilizes. For one of the first times in history, we created deepfakes of everyday people, rather than celebrities, using a limited amount of video and auditory data to build face and voice profiles from. The world sees deepfakes as dangerous and invasive; but Dove’s usage of the technology to ignite conversation around one of the largest issues girls face globally–toxic influencers and toxic beauty advice – reimagines it as a force for positive change and an opportunity for purposeful brand-building.

The harmful advice, trends and products promoted by toxic influencers are creating a self-esteem crisis among teenage girls: 92% want to change the way they look, and 1 in 2 girls follow an influencer that makes them feel less confident. Our brief was twofold: firstly, to highlight the scale and harm of toxic trends across social platforms for parents, who are often unaware of what their daughters are seeing; and secondly, to help girls recognize and avoid toxic advice online. Tragically, toxic advice has now become so common, girls don’t even realize it’s toxic anymore. Our goal was to create a single, shareable film that would be relatable for both parents and their teen daughters, which spotlighted the issue as well as educating parents about resources from the Dove Self-Esteem Project they could use to help their daughters detoxify their feed.

The creative idea was to use deepfake technology to put the words of toxic influencers – words heard every day by young girls – into the mouths of the one person they trust most in the world: their mom. During casting, we secretly captured data on the moms’ facial movements, expressions and speech patterns, enabling us to create high-quality deepfakes without the moms ever knowing. Toxic trends which were popular worldwide were chosen for the experiment, ensuring the film resonated globally. Deepfakes are seen as dangerous and invasive; as the first brand to use deepfakes for good, we legitimized the technology for purposeful commercial work and showed how a classic brand could modernize itself while tackling one of the most pressing problems facing girls today. Only launched in late April, the work was covered in industry press in over 60 countries, and went viral on three continents.

Sadly, the social media algorithms that feed young girls toxic beauty advice show parents completely different content on the same platforms. We needed to spotlight the issue of toxic advice to parents who are unaware of what their daughters see online. Our strategy began with an insight true to all parents: you would never let your daughter hear harmful things in her own home. But it happens on social media every day. By highlighting the ubiquity of toxic trends and influencers on social media, we could make it clear that toxic advice is prevalent and ensure parents didn’t feel shamed or accused of not doing enough. Showing the full extent of toxic advice meant we could truthfully depict young girls’ experiences on social platforms, while informing and educating their parents. In this way, we could create an actionist platform that inspired parents to start a conversation with their daughter.

The social experiment was constructed in a three-week timeframe from the completion of the casting process. Once the facial and auditory data had been gathered, we quickly worked to build five hyper-realistic deepfakes that looked and sounded exactly like the mothers. Utilizing comprehensive social listening ensured our representation of girls’ experience online was true-to-life and relatable. We tracked fifty toxic trends in real-time across multiple social platforms, inputting them into our deepfake models two days before the experiment to ensure relevance. Finally, with natural processing AI, we built transcripts for the deepfakes based on the most popular phrases of hundreds of toxic influencers, ensuring that the deepfake moms spoke just like the influencers their daughters follow. The campaign launched in the US, Canada, UK & Brazil in late April 2022; it will then roll out in fourteen other supporting markets across LATAM, EU and South Africa.

Immediately, the film was organically covered by media and television outlets across the world and went viral on three continents. In the first ten days, the campaign achieved 13.6 million organic views in the USA, UK, Canada and Brazil, 1.1 billion earned impressions in the USA alone, and has a 99% positive campaign sentiment, which is especially impressive given the toxicity of the subject matter and the use of controversial deepfake technology. The campaign has driven millions of parents to digital tools and a downloadable Confidence Kit on the Dove Self-Esteem Project website, where average dwell time in the week after the film launch increased by 100% to over four minutes. Our largest market (USA) has only run 2% of its paid media to date, with the campaign running until end August. It is currently on pace to become Dove’s most viewed and most successful purpose-led campaign in history.