Deepfakes have potential beyond causing harm, just avoid Liar’s dividend trap: Experts

Deepfakes, which leverage advanced AI to create hyper-realistic but false images, videos, and audio, are blurring the thin lines of reality and fiction in unprecedented ways. As these technologies proliferate, the average Joe faces the daunting task of discerning truth from falsehood, leading to the potential for widespread misinformation and a phenomenon known as the “Liar’s dividend” – where the mere existence of deepfakes undermines trust in genuine media.

The second part of this feature story explores how individuals can become more adept at identifying deepfakes and navigating this complex landscape. We also delve into the positive applications of this technology in marketing and advertising.

Also read:

Avatar 2.0: Is responsible use of deepfake technologies possible? - Part 1

Liar’s dividend

“In today’s digital world, identifying deepfakes and dealing with the “Liar’s dividend” – where genuine content is dismissed as fake – requires some know-how,” says Sumit Gupta Founder-CEO, Viral Pitch. As someone deeply involved in the digital and content marketing space, Gupta believes it starts with educating ourselves about the deepfake technology, using detection tools, and looking out for inconsistencies in videos. “Verifying sources and using educational resources are also important. Tools like Deepware and Deeptrace assist in analyzing content, while entities like Mozilla and MIT’s Media Lab provide helpful training materials. Staying informed and vigilant is key to protecting ourselves and our communities from the potential pitfalls of deepfakes,” Gupta asserts.

John Paite, Chief Creative Officer, India, Media.Monks, agrees that deepfakes indeed make it challenging to distinguish between real and fabricated content, potentially leading to the truth itself being questioned or dismissed.

“For anyone who is familiar with the tech can more often than not, tell the difference, but even then it could get quite challenging. One issue with AI today is consistency; tools are being developed, some still at the Minimal Viable Product (MVP) stage, that can analyze and detect inconsistencies in facial movements and speech, which may indicate manipulations. Educational platforms like Udemy provide courses on media literacy, and organisations like the News Literacy Project offer guidance and tips to help navigate this landscape. Ultimately, relying on credible sources is becoming more crucial than ever,” he adds.

“If you look closely,” states Aakash Goplani, Account Director, SoCheers, “you'll likely find that nine out of ten Deepfake videos have inconsistent lip sync, making it easier for most people to tell real content from fake. Also, watch for unnatural movements, poor lighting, and mispronounced words (like names, locations, multi-syllable words, etc.) as additional signs of inconsistencies. To clear any further doubts, use a reverse image search to verify the image’s origin. Always be critical of what you see online and check information from multiple sources. It is important to remember that anyone can create a deepfake, so always question the authenticity of what you see,” Goplani advises.

Rohan Naterwalla, Executive Creative Head, Punt Creative, predicts that AI avatars are only going to get better with time. “While they’ve become increasingly real – for the most part, one can intuitively tell that something isn’t kosher when exposed to them for more than a couple of seconds. However, this is the proverbial “first iPhone”. It’s already really good, but they’re only going to get better with time. Hence, now is when we as a society take tangible counter measures so that we don’t find ourselves living in a world we don’t like, further than we already do,” points out Naterwalla.

He suggests a self-declaration form! Starting by cleaning our own house first, adds Naterwalla, it only makes sense to make an amendment to the self-declaration forms that the ad and marketing teams have been expected to fill-in in recent weeks. According to him, there could easily be a section that requires brands to declare if something they’re releasing has elements of AI or in some cases entirely made with AI.

• Legal and Regulatory Frameworks:Governments and international bodies should develop comprehensive regulations that clearly define the acceptable use cases of deepfake technology.
• Technological Countermeasures:Investment in technology that can detect deepfakes is crucial. Whilst Instagram in recent months have made some attempts to label work “Made with AI” - for the most part users have been left feeling more can be desired.
• Ethical Training and Awareness:Companies utilizing deepfake technology should implement rigorous ethical training programs for their employees, emphasizing the potential consequences of misuse. Public awareness campaigns can educate the general populace about the existence and implications of deepfakes.

Dr Vikram Kumar, Founder-MD at SRV Media, suggests a complex strategy to identify deepfakes and navigate the “Liar’s dividend”, or the dissemination of misleading information:

1. Educate Yourself on Deepfake Indicators:Being able to recognize fake indicators such as odd blinking or inconsistent facial movements can go a long way in pointing out manipulated videos. Knowing this may mean the difference between what is real and what is not.
2. Invest in AI-Powered Verification Tools: AI-powered verification systems integrated with blockchain technology should be prioritized by companies so that they can authenticate digital assets and ensure that content remains original. Without these tools, trust would be destroyed.
3. Utilize Reverse Image Search:Enabling reverse image search tools on social media and news platforms could help users ensure the authenticity of images and videos quickly, and identify potential manipulations.
4. Foster Industry Collaboration:Social media platforms need to work closely together with tech companies so that detailed guidelines are created to monitor any potential cases.
5. Monitor and Protect Your Brand:Brands should monitor their digital presence very closely and react instantly to any misrepresentation that might occur from deepfake videos. Being proactive in developing effective strategies in order to avoid the “Liar’s dividend” trap is essential.

Positive applications

Beyond the dangers, deepfakes also offer potential for positive applications. This technology can be ethically and beneficially deployed, particularly in the fields of marketing and advertising, where it can enhance creativity and engagement.

Sumit Gupta points out that despite the risks, deepfakes have great potential for positive use, especially in marketing and advertising. According to him, they can facilitate personalized marketing messages, localize content without reshooting it, and improve educational and training tools. “On top of that, deepfakes can enable engaging entertainment experiences and make it easier for those with hearing or language difficulties to understand ideas in a given piece of work. When employed with morality, this kind of technology will be able to revolutionize our approach towards targeting audience groups, creating content and giving back to them through innovation thereby enhancing growth across different industries.”

“If I can speak from an Ad industry, perspective particularly when using deepfake, it is crucial for brands to be transparent. This helps maintain trust and integrity, ensuring that consumers are fully aware of the nature of the content they are viewing. With proper disclosure, deepfakes can significantly enhance personalized advertising experiences. For instance, it can be used to localize campaigns effectively, adapting messages seamlessly across different languages and cultural contexts with one actor. Additionally, deepfakes enable user-generated content (UGC), allowing consumers to create tailored versions of campaigns, which can make ads more engaging and relatable. They also offer a respectful way to recreate iconic figures for nostalgic marketing, providing tributes that resonate with audiences. There are more ways that can use deepfake but in the end, proper consent and transparency should be the first and last thing,” says John Paite.

Aakash Goplani is of the opinion that deepfakes have potential beyond causing harm. “We have all seen how they can create personalised marketing experiences tailored to individuals, as demonstrated in campaigns such as Sunfeast Dark Fantasy and Cadbury Diwali Campaign. Additionally, deepfakes aid accessibility by providing sign language interpretations or language translations, making the content accessible to a wider audience. They even aid education with realistic simulations and training scenarios, helping with better product and service demos. Lastly, deepfakes enable virtual try-ons, influencer collaborations, and engaging product demos. However, ethical use is key, with transparency and consent being the top priority,” Goplani concludes.

Rohan Naterwalla feels that the positive use cases of deepfake tech are boundless. “Just in the past year or so, brands like Cadbury and Zomato have demonstrated great work using deepfake technology to deliver hyper-customized communication to their audiences. For example, Zomato utilized this technology in pre-roll ads featuring Hrithik Roshan, who would address viewers and customize the message based on their location. Similarly, Cadbury effectively made Shah Rukh Khan the brand ambassador for all SMEs in India through personalized advertising campaigns,” Naterwalla adds.

He lists out some other innovative applications of this technology that advertisers might use in the coming weeks and months:

• Cost-effective content creation:Producing high-quality video content can be expensive and time-consuming. Deepfake technology can streamline this process by enabling the creation of realistic avatars and environments without the need for extensive resources, thus democratizing content production.
• Language localization:Deepfakes can be used to localize advertisements for different regions by seamlessly altering the language and cultural nuances of the content. This ensures that campaigns are more relatable and effective across diverse markets.
• Accessibility and inclusion:Deepfake technology can be used to create content that is accessible to individuals with disabilities. For example, generating sign language interpretations or providing descriptive video services can make advertisements more inclusive.

“If the infallible truth about humanity is that newer tech will always be adopted, then it is perhaps only wise for us to start building a world that is ready to make the most of its positive ramifications,” says Naterwalla.