Algorithms and data have a special relationship and the intermediaries between them, which form these relationships can, in due course, have technological backgrounds, which we see growing on a perspective basis. It is undoubtedly true that certain platforms are utilizing Customer Experience (‘CX’) and User Interface (‘UI’) operations to automate consumer-related news and at the same time, these tools have been utilized to counter the waves of fake news/misinformation. However, we must realize that the incremental development of the same has been based on certain false equivalences, which affect the transformation of the field of Artificial Intelligence (‘AI’) and media sciences itself. This article addresses one of these false equivalences – that of ideological conflicts.
The transformation of AI Ethics is not credited merely to the advancement of the information age and the advancement of computational intelligence and Machine Learning (‘ML’). It must be noted that the latent transformation of artificial intelligence has been related very importantly to the sciences related to business, economics and management. The models of emotional intelligence, leadership, vertical and horizontal hierarchies & cultural intelligence devised by scholars and organizations primarily in the United States for a long time, have been focusing on a Eurocentric vision towards human originality and integrity. However, at a civilizational level, the transformation of AI Ethics must not be merely based on the issue of data being biased, and consequently on the assumption this data bias leads to the biases at a policy level. Secondly, the role of technology should not be leapfrogged beyond an extension of humankind; doing so may undermine the global and national values of countries that we can see in the fourth stage of neoliberal globalization, such as the social protections available especially amidst the COVID-19 pandemic.
On that note, there are certain structural problems with the conception of AI Ethics, which cannot be ignored;
- The predicator that AI can transcend the integrity, existentiality and originality of human utility and actions – as is faked and often misrepresented by companies and NGOs so often – is unreasonable;
- Auditing, attesting and analysing the scope of any AI system can be based on a fluid concept of virtual horizontal hierarchies of the AI systems, where it depends where AI is a third party, a subject or an object to the subject-matter;
- The problems of Blackbox AI [[1]] and algorithmic biases are interlinked with trade secrets and twisting of ‘AI Ethics Boards/Technology Boards’ policies, which even at a multilateral level, cannot be achieved until the indigenous considerations of countries is affably achieved;
- Biases exist, but the strategy to harmonize depends on the art of calibration and consumption of the biases, which under a Eurocentric and Americanized vision [[2]], is impossible to be achieved;
The UNESCO AI Chair, John Shawe-Taylor, Director, CSML, commented in an interview that human biases and their inherence has a lot to do with AI-related biases and he focused on the contour of experience for humans via AI. Similarly, Professor Sandra Wachter from Oxford Internet Institute, co-authored an article on the non-possibility of automated fairness in the event of automated discrimination. The usage of AI will have a high-level impact in the field of media sciences and the industry at least in the United States will rise positively at least as of the predictions recently provided. Interestingly, narratives are being written that ‘neoliberal’ data-related companies have an anti-socialist bias in the US; a narrative which can be disproved by many examples, including the censorship of a New York Post article on Hunter Biden, the second son of former US Vice President Joe Biden. Many users on social media portal, Twitter, report that they were not able to post the article link on their feed and if they did share it, then they were even blocked. In fact, a report by a media research centre shows how Twitter and Facebook have censored the President of the United States’ social media accounts and the accounts belonging to his re-election campaign at least 65 times. In contrast, the said companies have not censored former Vice President and Democratic presidential nominee Joe Biden and his campaign accounts. Facebook spokesman Andy Stone, a former Democratic operative, said last week that the platform would reduce their distribution pending fact-checks. Now, the issue related to biases, particularly of social-media posts and videos can be raised affirmatively, which in no ways is unreasonable. However, acting in a politicized manner to affect the state of sharing content on the basis of ideological biases is a worse case of private censorship, which in the end is guided by algorithmic moderation.
This issue of algorithmic moderation based on ideological biases may affect elections in India, the UK and even other States where populism has been regarded as a phenomenon against certain establishment politicians. It must be understood that affirming rule of law in issues related to fact-checking must be based on a proper dissection of information, and should not be based on any obscuration on the basis of ideological biases. Instead, fact-checking and reducing the growth of misinformed content must be done on a non-partisan basis, where a better and deeper outreach of the political factions must be done, whether socialist, libertarian or conservative. Furthermore, we must also realize that algorithms cannot understand the concept of memes or other dynamic graphic or textual content, which:
- can commit to virtue-signalling;
- are limited to information-sharing but can be detrimental to a particular set of people, even if the reporting is factual and accurate, and way accessible; and
- are not exclusive and need not be limited to a set of understanding of misinformation and fake news education.
Furthermore, even in the recent QUAD 2020 summit, only India and Japan have yet not agreed with USA and Australia on combating misinformation, which is due to various issues related to information warfare.
The politics of misinformation by definition, can be based on a severe technocratic definitive towards issues of public and private interest. If algorithmic operations are incapable to positively discriminate between the natural occurrence of information warfare in geopolitics, and the normal and normative recourse of misinformation, then like the American model, unfortunately, Facebook and Twitter can censor content across the world because of the lack of consensus. Also, the recognition of the first principles over censorship and free speech requires clearance and clarity, and countries can stand up at multilateral forums to adopt plural approaches to the same. However, applying a Eurocentric approach towards fact-checking and tackling misinformation is impossible, and if attempted, would then cause an array of problems in the social, cultural, political and individual space of people, which affects both individual and collective rights, and can suppress the indigenous dignity and liberty of the political will of the people. Such biases can also drastically promote the notions of power over competence in democracies. For example, the blame affirmed by Donald J Trump over private censorship and a statement by Ravishankar Prasad to Mark Zuckerberg over Facebook’s anti-center-right bias are not partly unwarranted either. Former partisan members from the opposition parties in the US and India have been accused of joining social media companies and instigating issues related to algorithmic bias. Therefore, algorithmic ‘justice’ must be combatted, but through relevant government interventions to balance the tilt of ideological biases; something which the US Government has to do considering the enormous influence of their social media companies.
Furthermore, it must be noted that the impacts of social media discussed by the film ‘The Social Dilemma’ on Netflix affects everyone, whether a socialist, a libertarian or a conservative. It not only affects the first principles of free speech, but also is naturally delegated at the level of cyberspace to private actors, such as companies, which are non-elected, and earn a lot from the users who join the cyberspace for free. For instance, in 2018, Mark Zuckerberg, the founder of Facebook was grilled by Guy Verhofstadt MEP from Belgium. In fact, Bytedance, the company behind TikTok, also censors any anti-China content, which ridicules or mocks the People’s Republic of China. Thus, the role of algorithms in international law in a recent Discussion Paper by the author of this article, also shows that the role of information warfare, culture wars and the naturalization of the subjectification of information coupled by the immense level of cultural appropriation and cherry-picking in moderation, cannot be limited to the immaturity and incompetence of the algorithms, as it has virtual consequences at physical and individual levels too and hence, must be effectively addressed.
(This post has been written by Abhivardhan. He is the CEO of Internationalism and Founder of Indian Society of Artificial Intelligence)
References
- Blackbox AI is any Artificial Intelligence system who operations, inputs and structure are not known/visible to users or interested third parties. It is an impenetrable, closed AI system. ↑
-
When it comes to Eurocentrism and the Americanization of issues related to technology or any social, political, cultural or civil issue, both these terms are used together to signify that certain biases on the aforementioned issues are usually imposed on developing economies and other developed countries that do not subscribe to the European or American views on society, law, culture, economics and others. The best example of the same is the idea by the Bush Administration in the US – followed by Barack Obama to impose democracy in the Middle East, that unfortunately led to chaos. Another example can be the plagiarism in music and other audio-visual content created by India’s Bollywood filmmakers, artists and others, which can be found in several films, from ‘Yaadon Ki Baarat’ in the 1960s. Even the legal interpretation of constitutional values of global north (developed, mostly Western) countries when is imposed on developing post-colonial states without any introspection into the social and political realities and problems of the society can be Eurocentrism and Americanization. A classic example is where the judgments on gender rights and women empowerment are applied subjectively and yet the free speech standards set up by the US Supreme Court are not implemented in India. ↑