Fakes Acche Hai

Not long ago, a deepfake video of a popular film personality had caused a stir in the social media prompting the authorities to censure the creator. His intention was to create an out of the box media content to grab attention. The darker side of this story aside, a similar drive, to create distinctive content, causes the digital doppelgangers to hit the ad campaigns in a big way. The deepfake impact forms a curious study topic for management students (e.g. online MBA, distance MBA, online BBA, distance BBA).

Powering the Brands

The term “deepfake” was first coined in 2017 in order to bring out humour clips (e.g. face swapping).  The deepfakes’ marketing potential was discovered thereafter and now it is the hot pick of admen. In India, deepfakes first appeared as product endorsement by celebrities (e.g. Sachin Tendulkar’s digital double endorsing an insurance company) There are other examples too which feature actors, Shahrukh Khan (Cadbury ad), Hrithik Roshan (Zomato ad) Salman Khan (Pepsi ad) and the like. Digitally impersonated political leaders have also been used in elections campaigns lately.

Deep + Fake = Deepfake

Deepfakes use highly sophisticated AI technology. As the name indicates, deepfake combines ‘deep’ (deep study) and ‘fake’ (manipulated images). Using the AI technology, features of an object (e.g. human face) is studied in depth from various angles and a look alike image, with incredible realism, is created and super imposed on the original.

The realism of the output depends on the quality and variety of the dataset of the object being mimicked (images / videos etc.) The richer the data set, the more realistic the deepfake becomes. (e.g. there are many online images of popular actors, hence their deepfakes appear more realistic).

The advanced algorithms used in deepfakes have the capabilities to revolutionize the application videos for entertainment, education, and personal communication (e.g. human face to chatbots or online shopping assistants).

The Plus Side

Beneficial Impacts of deepfakes are many, as it:

• Facilitates cost effective branding solution (e.g. Celebrity endorsement at low cost)
• Enables repurposing (e.g. Shah Rukh Khan’s “Not Just a Cadbury Ad” being used by local retailers)
• Adds variety (e.g. Dhoni singing Kishore Kumar song in an e Bike company ad)
• Supports personalization (e.g. Hrithik Roshan speaking about select local food outlets in the Zomato ad)

Minuses Too

Any resource can be put to good use and misused as well. Deepfakes are no exceptions. We have come across instances of deepfaked endorsements/impersonations to create disharmony, financial frauds, etc. Similarly, fake product endorsement by celebrities is another menace. Awareness initiatives by the authorities do work to some extent.

Is Spotting Feasible?

In the context of deepfake misuse posing security and ethical challenges, experts suggest a few basic tips to mark off the deepfakes from the original:

• Eye Movements (e.g. eye blinking may not look natural in deepfakes)
• Asynchronous facial expressions vis-a-vis emotions.
• Lip movement mismatch with spoken words.
• Voice discord (does not tally the voice quality of a known personality)

But the disclaimer is that they are not time-tested and may fail in the ever changing threat landscape.  The remedy is to constantly explore and innovate the defense.

The Final Word

Staying relevant to latest marketing technology is essential to make the brand appealing to the tech-savvy audience. The misuses notwithstanding, the power of well-executed deepfakes in marketing campaigns just can’t be ignored. Harnessing their potential would help you to gain the required niche. At the same time companies need to exercise due diligence that the synthetic media do not infringe the mimicked person’s privacy.  They also be prudent to protect the customers against any abuse using deepfake technology.