Uncovering Deepfake Technology
BY CHESTA CHOPRA/ NOVEMBER 26, 2022
Technology today offers both frightening powers and creative possibilities. The development of deepfakes has made expensive VFX techniques more accessible. The quality of videos gets better with lower costs but, what if it costs us our ability to differentiate between real and fake?
n the present era of changes, brands are constantly striving to make their customers feel special. One such instance was when a commercial advertisement for Cadbury's Raksha Bandhan celebrations featured Hrithik Roshan personally wishing viewers a Happy Raksha Bandhan. This was employed by Cadbury for both Raksha Bandhan and a Diwali campaign in which Shah Rukh Khan supported small and local businesses. In contrast to large companies, small businesses in India lack access to superstar brand ambassadors, Cadbury claimed that this was the best way for them to aid in their recovery from Covid. By utilizing the machine learning algorithm and the viewer's pincodes, various versions of the same advertisement were made with local store names. In this case, Cadbury employed a machine learning technology to replicate Shah Rukh Khan's visage and voice. This new machine learning technology is known as ‘Deepfakes'.
I
A deepfake is an Artificial Intelligence generated video or audio clip of a real person doing and/or saying real/fictional/imaginary things.
A deepfake is an Artificial Intelligence generated video or audio clip of a real person doing and/or saying real/fictional/imaginary things. It uses generative adversarial networks or GANs, a type of generative modeling based on deep learning or artificial intelligence, to superimpose existing video footage of a face onto a source head or body. They are called deepfakes because they use deep learning technology, a branch of machine learning in which, when given enough data, a computer can produce output on its own and thereby produce such fake images and videos. As a result, we say that they are digitally altered. This article explores and examines the economic impact of deepfakes on politics, democracy and above all society though the major use of deepfakes lies in the illicit video making sector.
.jpg)
Illustration by Unknown via Pinterest
The central argument against deepfakes is its contribution in blurring the lines between what is real and fake.
The central argument against deepfakes is its contribution in blurring the lines between what is real and fake. If seeing is no longer believing, the weakening influence on democracy will take on greater importance. As Abraham Lincoln said, ‘Democracy is for the people, by the people, and to the people’, but it won't function if the people simply don't believe in it, and start losing their trust in this system. While democracy is often associated with higher human capital accumulation, lower levels of inflation, lower political instability, and higher economic and political freedom, if democracy's base is even marginally undermined, the entire system would crumble. Moreover, as technology develops, deepfake could potentially be used to deceive voters in democracies. The issue here is that when we lack the ability to distinguish between what is genuine and what is not, the propensity to make poor choices grows. When the ability to make intelligent decisions is diminished, the precision of the cost-benefit analysis that a customer faces is also reduced. As a result, the economy suffers.
.jpg)
Illustration by Unknown via Pinterest
On the political front, the technology first came to light in 2018, when a comedian imitating Barack Obama released a video discussing the dangers of deepfakes. According to the Washington Post, the president of Gabon, an African country was reportedly featured in a suspected deepfake video. The country's military even attempted a coup because they thought something was amiss in the video. In another incident, Bill Poster created a Mark Zuckerberg impersonation video to gauge internet’s reaction over computational propagandas, but even after the firm demoted it, the video continued to go viral. Both these examples demonstrate the potential for deepfakes to spread misinformation online. Spreading false information
online has an impact on both our trust, and our pockets. According to an Israel-based cybersecurity firm CHEQ and the University of Baltimore, fake news is costing the global economy USD 78 billion each year. The study also reveals that fake news has contributed USD 39 billion annual loss in stock market value of 9 billion a year. When text, pictures, and even low-tech fake videos can have such a significant impact on the economy, the dangers of deepfake are incomprehensible. Deepfakes may possibly be utilized as a defense to shield politicians from criticism. It will allow corrupt politicians to contort information and avoid accountability. One of the most threatening consequences of this may be the increasing number of individuals who choose to stay away from politics, which will result in ‘reality apathy’.
When text, pictures, and even low-tech fake videos can have such a significant impact on the economy, the dangers of deepfake are incomprehensible.
​On the contrary, according to the Quint, “In 2020, a video of politician Manoj Tiwari was released with the help of deepfake to reach a larger voter base.” While his original video was in Hindi, similar videos were created in Haryanvi and English using this technology. According to The Vice, Neelkant Bakshi, co-incharge of social media and IT for BJP Delhi, said that deepfake technology has helped them in scaling their campaigns. They were able to target different sets of audiences through three different languages. The housewives in Delhi were overjoyed seeing their leader speak in their language. Due to this, many people are now adopting deepfake. These two instances show how strategically using deepfakes can have a positive effect on the economy and society, it helps in reducing the cost of production and thus increase profits and efficiency.
Lower costs and subsequent benefits are especially beneficial for the film industry as they can use it to update film footage rather than reshoot it, which will lower the cost of making a movie.
​
Lower costs and subsequent benefits are especially beneficial for the film industry as they can use it to update film footage rather than reshoot it, which will lower the cost of making a movie. Voice dubbing will also sound more realistic by using deepfakes. For instance, in a 2019 campaign to raise awareness of malaria, David Beckham featured to be speaking in a variety of languages and addressing a variety of audiences.
Deepfake has the ability to make expensive VFX techniques more accessible. For a small fraction of the price, it can also develop into a potent tool for independent storytellers. Synthetic media can help journalists and activists for human rights maintain their anonymity under repressive and autocratic governments. By using this technology to report atrocities on conventional or social media, journalists and activists can potentially acquire significant power.
.jpg)
Illustration by Unknown via Pinterest
To safeguard people's privacy, deepfake can also be used to spoof people's voices and faces. People can utilize Deepfakes to make avatar experiences for online self-expression. By using a personal digital avatar, individuals can increase their autonomy and broaden their purpose, thoughts, and values. People with physical or mental limitations will be able to express themselves through artificial avatars online.
It is also alarming that according to a research conducted by Iproov, only 29% of people in India are aware about deepfakes in 2022.
Research by a UK based think tank revealed that deepfake videos have doubled from 13% in 2019 to 29% in 2022. However, to deploy deepfakes in the modern world, we need more knowledgeable people who are familiar with this technology as these videos may be distinguished from real ones using machine learning. In India, this would lead to the creation of employment possibilities. India has a low level of digital literacy with more than 90% of its population being digitally illiterate (Modi D, 2022). It is also alarming that according to a research conducted by Iproov, only 29% of people in India are aware about deepfakes in 2022. Due to the lower levels of digital literacy and awareness when even low-tech fake videos have led to violence, the ability to critically identify a deepfake video becomes crucial making it possible for people to avoid falling for its tricks and propaganda techniques. Poor-quality deepfakes are easier to spot. Fine elements like lip-syncing, skin tone, and hair quality are particularly difficult for deep fakes to simulate with present technology. Monash University and IIT-Ropar have created a unique detector ‘FakeBuster’, which aids in the detection of deepfake.
_edited.png)
Illustration by Unknown via Pinterest
If society wants to remain a step ahead, the government and private educational institutions need to fund deepfake detective technologies. However, there is no explicit law in India prohibiting deepfakes. The technology might not be advanced right now, but will be so in the future and this is why due importance needs to be paid to it. Hao Li, who is an expert in deepfake technology, says that “This is developing more rapidly than I thought. Soon it’s going to get to the point where there is no way that we can actually detect it, so we have to look at other types of solutions’’. Today, content creators compete for audience' attention on social media platforms including Facebook, TikTok, and Reddit. Regular visitors to these video-sharing websites are becoming more and more aware of deceptive videos that content creators create. They are inclined to create emotional or startling videos due to strong financial incentives. As a result, they stage dramatic or disturbing scenarios and frequent viewers are able to identify these staged videos. As deepfake videos proliferate more, people will eventually become more aware that they shouldn't always accept what they see.
​
Keywords
deepfake, machine learning ,GANs , democracy, cost-benefit, reality apathy, synthetic media , VFX
​
References
​
Volkert, R. Deep Fakes Understanding the illicit economy for synthetic media, NISOS.
https://cdn2.hubspot.net/hubfs/6068438/Resources/NISOS%20-%20Deep%20Fakes%20White%20Paper.pdf
Simmons, J, R (2020, May 13). The Real Business Impacts of Deepfakes, Nojitter.
https://www.nojitter.com/security/real-business-impacts-deepfakes?ng_gateway_return=true&full=true
Brown,E (2019, December 18). Online fake news is costing us $78 billion globally each year, Z D Net.
https://www.zdnet.com/article/online-fake-news-costing-us-78-billion-globally-each-year
Bearak,M (2019, January 7). Gabon foils coup attempt against absent president. Washington Post.
(2020, February 19). Delhi Polls: Manoj Tiwari Used Deepfake to Reach Larger Voter Base, The Quint.
Christopher,N. (2020, February 18). We've Just Seen the First Use of Deepfakes in an Indian Election Campaign. The Vice.
https://www.vice.com/en/article/jgedjb/the-first-use-of-deepfakes-in-indian-election-by-bjp
How deepfakes could change the internet (2021, November 27) Youtube.
https://www.youtube.com/watch?v=o7J8I7ag1nE
(2022, August 26). How To Protect Against Deepfakes-Statistics and Solutions. Iproov.
Sample,I (2020, January 13). What are deepfakes – and how can you spot them?. The Guardian.
https://www.theguardian.com/technology/2020/jan/13/what-are-deepfakes-and-how-can-you-spot-them
Wigmore, I. deepfake AI (deep fake). Tech Target.
DISCOVER MORE STORIES
NOVEMBER 19
​
BY YASHODHARA KATAKI
NOVEMBER 12
​
BY GRACE D'SOUZA
The Arab Israeli Conflict: An Indian Standpoint
NOVEMBER 5
​
BY BHAVYA SINGH
The views published in this journal are those of the individual author/s and do not necessarily reflect the position or policy of the team behind Beyond Margins, or the Department of Economics of Sophia College for Women (Autonomous), or Sophia College for Women (Autonomous) in general. The list of sources may not be exhaustive. If you’d like to have the complete list, email us at beyondmarginssophia@gmail.com