User:Rifat mohd/sandbox

= Contribution A =

Hyper-reality and Synthetic Media
The term ‘hyper-reality’ was first used by Jean Baudrillard to distinguish simulated reality from actual reality. According to Baudrillard, people are drawn to aesthetic content regardless they are from an original source or a simulated one. This thought stands true to the rise of synthetic media in the current world. Synthetic media refers to any media which is created or enhanced with the help of algorithmic automation. Recent technological advancements in artificial intelligence and the democratization of new technologies have made the creation of synthetic media easily accessible for the masses. Examples of synthetic media include deepfake, virtual influencers, and synthetic images such as computer generated imagery. Portrait mode in iPhones stands as a good example for synthetic media where photos are manipulated to look like they were taken from a high-end camera.

Recent advancements in artificial intelligence and algorithms have made it increasingly harder for the audience to distinguish hyper-realistic media from reality. Deepfakes contributed the most to the distribution of synthetic media. They mainly rely on deep neural networks and AI to create content by working on a source image. Hence, they further remove the need for artistic input by depending on algorithms and image data fed into the autoencoder. Popular Chinese application, Zao is a good example of a deepfake application. It allows users to star in any movies or television shows by using face-swapping deepfake tools. According to Kietzmann, Mills, and Plangger, the underlying algorithms of deepfakes are advancing quickly to a point where users in the near future would be able to create high quality commercials from it.

Authenticity of Deepfake Content
People have adapted to the idea of image manipulation and there is a skepticism applied to photos circulated in media without approvable sources. However, the same cannot be applied to video graphic content where people readily accept it. This makes deepfake technology a potential candidate to create alternate realities regarding the information on politics, individuals, and brands. Hence, it is an important responsibility of brands to acknowledge their use of synthetic media in their content whether it is a source of entertainment or providing a subscription service - which acknowledges its use all the more important. The financial fraud of a UK-based energy company using audio deepfake sets a prime example for one of its misuse. One of the employees received a deepfaked call claiming to be the CEO of the company, asking for funds for a business deal. This lead to a financial fraud of €220,000 which was later found to be directed to Mexico. Hence, such easy accessibility of it could make scammers more menacing by using audio and video deepfakes to sell their illusion.

Therefore, it is particularly important to invest in technologies that detect deepfakes. According to Hasan and Salah, blockchain serves as a foolproof method to find the original source of the content. Due to its tamperproof records, logs, and open public transactions, blockchains would be able to find the original source of content before it is simulated or replicated with false information. Maras and Alexandrou identify that the research on detecting manipulation in video content is scarce. Moreover, they were able to identify that the same algorithm used for face swaps could be used to detect whether a video is deepfaked or not. Furthermore, the United States Defense Advanced Research Projects Agency (DARPA) took the initiative to develop technologies that automatically detect altered video contents.

= Contribution B =

Ethical Issues of Artificial Intelligence
In this digital era, machine learning is utilized to gather sensitive information of the citizens such as their sexual orientation, interests, livelihood, and life expectancy. There have been reports of the Chinese government using deep learning networks to identify and track the Uighurs, a minority ethnic group in China. One of the main issues identified using AI is that the information gathered is biased and unethical. Machine-learning algorithms are found to be scripted with discriminations such as sexism and racism which is affecting how they gather and present sensitive data in our society. According to Reuters, Amazon relied on AI to sort through resumes for hiring new employees. This resulted in the AI neglecting the female candidates or anyone who attended a women's college. Moreover, Nikon's camera AI is found to misread Asian people as blinking and also HP's webcam software's face localization algorithm had a difficult time tracking people with dark skin tones.

The underlying issue is because of the nature of data fed into the algorithm for machine learning. AI learns by analyzing data sets fed into it by the chosen engineers which further proceeds to build a system model based on those data. Hence, if a system is fed more images consisting of white males, the AI will have a difficult time recognizing dark-skinned women. According to Boyd, Levy, and Marwick, algorithmic means of decision-making will accelerate discrimination based on an individual's personal network since there are no existing rules to stop it. They further observe that the discrimination laws are currently solely focused on an individual or a group where instead it should also include their positions within the network. Prejudice, sexism, and different types of extremism and bias are as yet unavoidable in contemporary society, yet innovations can be baked in with such discriminations in the absence of a proper ethical model for networks.

Major companies such as Microsoft and Google are already investing their time to build a model for AI that addresses the ethical use of data. Microsoft has established 6 principles - reliability & safety, privacy & security, transparency, accountability, fairness and, inclusiveness for AIs for data collection. Likewise, Facebook has invested in the Technical University of Munich to start up an institute that specializes in ethics for AI. Moreover, the G20 summit of 2019 came up with its list of principles for the usage of AI that respects human rights. Another advantage of the algorithmic approach is its robust distribution and updatable nature. Compared to other technologies, once implemented, algorithms can be easily reconfigured to fix a certain issue via a software update. The HP's face localization issue was fixed via a software update that was deployed within a few weeks of identifying the problem. Hence, a modular and transparent approach to creating these algorithms can minimize such ethical issues.

= Response to My Peer's (Charan Siddarth) Contribution A   [Written by Rifat] =

Authenticity of Citizen Journalism
Professional journalists use aesthetics and ethical principles to distinguish themselves from citizen journalists. However, according to the research by Mortensen in 2014, the photo quality of citizen journalists was not far off compared to the professionals. Similarly, according to the research by Puustinen and Seppanen, amateur images from the citizens were found to be equally trustworthy or sometimes even more trusted than the professional photographs. This could be due to how we perceive amateur accidental photography more closer to realism compared to the carefully framed professional ones. Calling back to Baudrillard's views on hyper-reality, we have reached a stage where the audience is finding it harder to distinguish reality from the simulated. Hence, this could stand as a testimony to Baudrillard's theory. Furthermore, AI algorithms have a major role in helping an average user capture close to professional photographs. By training their deep neural networks, companies like Google have a solid algorithm that helps them to identify the photos captured by the user and automatically label and stylize them with the help of AI even on low-powered devices. Likewise, the rise of synthetic media and its easy accessibility to the general public has led to a tremendous increase in fake content creation and distribution. Interestingly, according to Shen, the general public categorized a particular photo to be trustworthy depending on the individuals technical skills, photo editing skills and their social media handles. He also added that the social cues such as source trustworthiness, bandwagon, and intermediary trustworthiness had no significant impact on the general public to dismiss a particular fake image.

= Response to My Peer's (Charan Siddarth) Contribution B  [Written by Rifat] =

Market Value of Personal Branding
Digitization has made personal branding a tool that is not just limited to companies and organizations but also individuals and their social lives. Branding is not just limited to exploring the visual identity but also to the experiences the individual goes through and the relationships they nurture. As Marwick states, “Self-branding is primarily a series of marketing strategies applied to the individual. It is a set of practices and a mindset, a way of thinking about the self as a saleable commodity that can tempt a potential employer”. Compared to 2 decades ago, consumers currently are also interested in the personality portrayed by the brand and it plays a major role in their consumption habits. Due to this fact, there is a trend among companies to invest in social media influencers since the personality portrayed by them resonates with the brand's personality. Hence, personal branding has a huge potential to shape and control the market. McCorquodale supports this by adding that people who are successful in engagingly portraying their own life has a better chance to be termed as an influencer which the brands are actively looking for to lure in their potential viewers. Hence, with the emergence of social media and micro-influencers, personal branding is an important step for an individual since it can affect their social life and even their employability.

= References =