There was palpable curiosity on the internet when a Deep fake first made its appearance in 1997. 

Wait, what are deep fakes?

It’s a form of synthetic media generated using Artificial Intelligence (AI) and deep learning techniques. This technology uses deep neural networks to analyze and replicate patterns from vast data sets, enabling it to swap faces in videos, clone voices, manipulate video content, and generate text that mimics a specific style. 

‍

‍

The excitement, however, quickly devolved into fear, when deep fake technology was utilized to create fake videos of different celebrities and politicians. Anybody with access to a neural network AI, could potentially spread misinformation and cause real-world issues, through misuse, deception and privacy violations.

‍

‍

However, brands and marketers are now exploring its potential use for creating compelling, hyper-personalized narratives, especially with AI tools maturing and finding applications in entertainment. 

‍

Deep Fakes - The Great, Good and Ugly

We analyzed three case studies on brands and deep fakes, exploring whether it’s a strategy for success. 

Here are some key takeaways:

• Deep fake technology can create an emotional impact and spark conversations on important issues.
• Personalization, driven by deepfake technology, has the potential to deeply engage consumers
• Brand authenticity and consumer trust can take a hit, when confronted with scandals and disputes such as the Bruce Williis case.  
• These kinds of campaigns, where issues of consent are murky, can sow doubt about the intersection of AI marketing and celebrity image rights.

‍

Case 1 - Great - Dove‍

‍

‍

Problem:

Unilever-owned Dove, renowned for its commitment to promoting body positivity and self-esteem, embarked on a mission to unmask the pervasive toxic beauty advice prevalent on social media.

‍

Approach:

• Dove's #DetoxYourFeed campaign cleverly harnessed deep fakes to create a thought-provoking and relatable narrative. At the heart of this initiative was the impactful film titled 'Toxic Influence.' 
• Using sophisticated face-mapping technology and AI tools, Dove literally placed the voices of influencers who promote toxic beauty advice into the mouths of mothers. Shock and disbelief washed over the families as the mothers uttered phrases such as "Botox is amazing, you're never too young to start" and "if your teeth are uneven, you can always file them down with a nail file." 
• Dove collaborated with Ogilvy, a renowned advertising agency that has made a bold commitment not to work with influencers who digitally altered themselves for advertisements, underscoring this message with thoughtful actions.  

‍

Results:

• Generated substantial social media attention.
• Elicited strong emotional reactions, highlighting the urgency of addressing harmful online content. 
• 72% of girls felt better after unfollowing idealized beauty content, according to a Dove report after the campaign.

‍

Learning:

• Deep fake technology can create an emotional impact and spark conversations on important issues.
• Collaborating with influencers and advocates enhances campaign reach and impact.

‍

Case 2 - Good - Messi and Lays

‍

‍

Problem:

In a season beset by visuals of empty stadiums, Lay's sought to ignite match-day excitement by using deepfake technology and their brand ambassador, Lionel Messi, to create a personalized web experience known as "Lay's Messi Messages."

‍

Approach:

• The brand used AI tools and machine learning to craft video messages tailored to individual users and their specific markets, languages. To make this experience possible, a brief shoot with Messi captured phrases and basic movements that were essential to the campaign videos. 
• Users could then create their own personalized video messages from Lionel Messi, choosing between ten different languages to generate a diverse range of bespoke invitations - all with corresponding lip movement. 
• Messi appeared on the users' screens, extending a personal invitation to watch the game, complete with references to their submitted names, dates, and locations. 

‍

Results: 

• The microsite garnered a staggering 38 million hits within the first 24 hours of launch, showcasing the campaign's exceptional popularity. 
• Messi shared the experience with his Instagram following of over 192 million followers, triggering a cascade of Messi Messages across social media platforms and sparking conversations worldwide.
• The AI marketing campaign created a remote brand experience with global shareability and forged deeper connections with consumers through personalization.  

‍

Learning:

• Personalization, driven by deepfake technology, has the potential to deeply engage consumers, creating a sense of authenticity in consumer interactions with celebrities. 
• Multi-language accessibility and the ability to share personalized content across borders demonstrates the global reach and virality potential of personalized brand experiences.
• Leveraging the influence of brand ambassadors and sharing experiences on social media can exponentially increase brand visibility and engagement.

‍

Case 3 - Ugly - Bruce Willis

‍

‍

Problem: 

A unique and controversial case emerged involving Hollywood icon Bruce Willis and deep fake company Deepcake. Reports circulated that Willis had sold the rights to his own face for a Russian Telecom company ad - a groundbreaking development in the realm of AI-driven celebrity image rights. But did he?

‍

Implementation:

Various media outlets, including the Daily Mail and the Telegraph, published sensational reports suggesting that Bruce Willis had entered into an unprecedented partnership with Deepcake - an AI-focused company that specializes in deep fakes. According to these stories, Willis had purportedly sold his image rights, allowing for the creation of a 'digital twin' for film and ad use. A deepfake of the actor was used in an ad for Megafon - a Russian telecoms company. 

Deepcake disclosed that they worked with Willis’ team on the project, adding that the actor had provided his consent and access to relevant materials to create his digital twin. However, this was not the case.

‍

Result:

In response to the reports, representatives for Bruce Willis categorically denied Deepcake’s claims, casting doubts on the legitimacy of Deepcake’s actions. The company responded through a statement refuting the reports, and emphasizing that Willis owned his image rights and that their purchase was inaccurate. The back-and-forth muddied the waters of this collaboration and tainted the campaign. 

‍

Learning:

• Brand authenticity can take a hit when confronted with scandals and disputes such as this one. 
• The evolving and ambiguous nature of the deepfake industry highlights the need for clearer regulations in the industry.
• These kinds of campaigns, where issues of consent are murky, can spark debates and doubts about the intersection of AI and celebrity image rights.
• The denial of any partnership agreement cast doubt on the ethics of deep fake companies. 

‍

And questions still linger for brands when using AI tools, 

• Can brands utilize this deep fake technology in a compassionate, ethical and transformative way? 
• Can they be transparent about the use of this technology? 
• And will deep fakes affect real change, or will the results remain skin deep? 

‍

What do you think about AI in marketing?

Should brands be using deep fake technology? Write to us.

‍