Misuse of technologies like deepfake is known to cause much harm. But do they have any advantages at all? When put to right use, the artificial intelligence (AI)-based technology can help scale up, save costs and increase engagement.
Elevate Your Tech Prowess with High-Value Skill CoursesSmall wonder then consumer brands such as Mondelez, ITC and Zomato are using deepfake tech for innovative advertising which allows viewers to co-star and collaborate with celebrities such as Shahrukh Khan, Hrithik Roshan and Sachin Tendulkar. Here’s a glimpse into how brands are putting the controversial tech to good use.
Mondelez won multiple titles at this year’s Cannes Lions Festival of Creativity – including India’s first-ever Titanium Lion for its AI-made Cadbury ad. The campaign allowed local shop owners to feature alongside actor Shahrukh Khan’s deepfake and create a free personalised ad for their small businesses by sharing basic information about their services. AI was then used to create a deepfake using the actor’s face and voice to talk about the store.
Zomato used a deepfake ad showing actor Hrithik Roshan craving specific dishes from popular restaurants in different cities. The food delivery company identified top dishes and restaurants in select pin codes which were detected using the phone’s GPS of a viewer.
ITC recently partnered with AI creatives company Akool to launch the #HarDilKiFantasy campaign for its biscuit brand Sunfeast Dark Fantasy where participants could share screen with Shahrukh Khan.
Pepsi created a deepfake of actor Salman Khan’s 1989 character Prem and featured him against his current self to build the ad narrative. Similarly, Ageas Federal Life Insurance created a deepfake 11-year-old version of cricketer Sachin Tendulkar and created a conversation with his adult self.
What is a deepfake video and how is it created?
A deepfake refers to synthetic (fake) media created using deep learning. These are different from ‘cheapfakes’ or ‘shallowfakes’ which are simple manipulation of an image or video by speeding, slowing or swapping faces, as in a Snapchat dog filter. Deepfake is an evolved AI methodology which does lip synching, puppeteering full body movements and voice cloning to create synthetic avatars of existing or non-existing persons.
How does it help marketers?
A deepfake advertisement has immense self-distribution potential because the audience directly engages with the campaign. The hyper-personalised content can go viral on social media within minutes. An average 15-minute shoot is used to create the celebrity avatar along with training the model on some audio and video data. Companies can even do away with the need of hiring a celebrity. For instance, the second leg of the Cadbury ad does not feature any popular name.
What legal protections are required for creating a deepfake ad?
A person or an entity can be held liable for infringing another person’s privacy or publicity rights for creating synthetic media. Using deepfakes maliciously could also constitute fraud, defamation, identity theft and other civil and criminal violations. Therefore, companies exercise due diligence while licensing images or videos of persons for commercial use.
Experts say these advertising contracts address minutest specifics such as depicting the actor in certain clothes or age or using his or her voice to utter certain controversial catchphrases.
For instance, Zomato had to pull down the Zomato ad in Ujjain where Hrithik Roshan was seen using a restaurant’s name similar to a temple’s name.
Zomato put out a public apology for unintentionally hurting the religious sentiments of the temple’s priests. Use of deepfakes in commercial advertising is an extremely sensitive legal issue because publishers do not have control over where the video goes and what content is created.