Reading Responses!

Looking for my reading responses?

You're in the right place! Unless... you're looking for my commentary or my home page. If this sounds like you, click on their respective links below:

My Home Page

My Individual Commentary

Reading Response #1

Response to "A Declaration of the Independence of Cyberspace"

- To what extent did Barlow’s vision turn out to be true? To what extent is it still relevant today?

- How does the Internet bypass (or not) state authority? What about other authorities?

The Cyberspace faces the same dilemma that international relations face, I write, as I prepare this response paper in the midst of an International Relations lecture. At the global level, states face absolute anarchy with no real authoritative power applying to enforce any specific rules or regulations. Cyberspace is the same — a post may be posted within the borders of one country but it may directly impact one or more people even on the other side of the planet.. Thus, Barlow’s assertion that there will always be attempts to control this international Cyberspace is true, and is still relevant today. The question becomes to what extent do and should individual governments have responsibilities over content and actors of the global space. Further, if it is not governments which must exert control and enforce stability, who (if any) can or should it be? Who is the “we “ to which Barlow so often refers, more specifically than just international users?

In many ways, Barlow frequents many elements from traditional declarations of independence and deven uses the classic metaphors of self-definition that dominate constitution-writing circles. He essentially suggests that “We the People '' should take over the decision making process, but falls into the same trap that constitution writers and elite lawmakers have created and fallen themselves into. It is difficult to define that “We” even in an ethnically homogenous nation, let alone a politically, religiously, ethnically diverse community transcending national borders. It is easy to shut out governmental regulation arguing for a new Social Contract, but how will anything be regulated (if it must be) in the process of defining that “We” and creating the method by which the Social Contract will be created and implemented?

Thus we are faced with multiple questions very similar to those that arise in the process of organizing a new nation. The only way Barlow’s Declaration could take legitimacy is if he were to ask and answer all these questions, ultimately leading to an international yet nation-like community on the Cyberspace, sovereign and separate from nation-states. This would require him to go deeper into his idea of a Social Contract, and to me, it reminds me of work we’ve done in my Constitutionalism in MENA class: It’s extremely interesting to think about the ways in which we could imagine constitutionalism in the Cyberspace. Andras Sajo and Renata Uitz define constitutionalism as a “set of interrelated concepts, principles, and practices of organizing, and thereby limiting power in order to prevent despotism.” Taking this approach to constitutionalism, it would be difficult to realistically hear anyone’s opposition against limiting overreaching power in the online realm: whether that's a government or a monopolistic platform.

In this day and age, the internet already bypasses state authority to a great extent, but not too differently from the way black markets of drugs or foreign currencies have and will continue to evade the law throughout the course of history. So Barlow actually brings up a really good point about the need to organize the Cyberspace similarly to the way in which we have organized our nations and our communities. The one distinction he makes is that it is impossible to nationalize the Space, and that we must therefore find an organization unique to it.

This summer I did an internship at the International Press Institute, and one of the many topics covered among our activism was the safety of journalists online and how states, but also GAFAM and other media giants could combat harassment and threats on their platforms. Quite often I saw countless experiences where tech platforms didn’t do enough to protect individuals against racist, sexist, etc. attacks, yet how, at times, their Community Guidelines” worked more so in favor of fascists spreading hateful rhetoric instead. Today’s reality thus begs the question: how do we take this power and create a “We” that takes power from nation states but also avoids despotic power imposed by GAFAM, and gives this power instead to average internet users?

Reading Response #2

Response to "Engineering the Public: Big data, Surveillance and computational politics"

- How do you think fake news existed before the Internet? And how did the Internet affect them?

- How have pre-existing systems of public manipulation been affected by networked digital media?

Both fake news and public manipulation existed before the internet. Tüfekçi’s article focuses particularly on the second aspect and the internet has both combatted and worsened public manipulation, but applying her framework to the spread of fake news might be beneficial as well. In this reading response, I’ll first share my personal thoughts on the new “public sphere” as described by Tüfekçi, and then consider how her analysis of big data impacts today’s fake news.

Tüfekçi puts forward the concept of “computational politics,” as an individualized, marketing-like approach to politics, which combines digital resources and behavioral sciences to individualize policies and campaign approaches. The impacts of this come in pros and cons: while the individual might be better served due to the vast amount of data about their preferences and behaviours, they may also be better manipulated and controlled thanks to the very same methods. Yet Tüfekçi’s findings on politics have similar impacts on the spread of news. Just as Tüfekçi’s six factors explaining the rise of computational politics may both benefit and damage campaigning methods and policy formulation, they may also lead to either more relevant, personalized, or more misleading, partisan news.

Before social media connected people across the globe, fake news was usually only spread about things that people could potentially be impacted — at a local scale. Conspiracy theories existed, but they were much more marginal and gained little traction locally. Now that the globe has connected extremists of all nationalities, we’re seeing that fake news is now spread by people who may have no impact or exposure whatsoever to the new that they are spreading. Whereas tabloid news still came in the form of articles, and took some time to read and digest, now we’re seeing flashy Canva headlines on instagram or 280-character tweets as some people’s only source for their information. This means that people mindlessly engage and spread news that may or may not be fake, whether or not it affects them or others.

In terms of Tüfekçi’s six variables, all of them also directly link to the increased spread of fake news. Big data allows for platforms and news providers to collect specific information about what kind of news people need/want/will actually engage in and perhaps spread. Computational modeling and behaviour facilitate this process of news aggregation and tailoring, while algorithms allow for the ways in which news is aggregate and recommended to remain opaque, making people really believe in the “God-like machine” recommending them the “best information possible.”

People are thus less likely to question both the source and the accuracy of the news they receive on their feed, because it goes through such an independent “black-box” as described by Tüfekçi. While people had some sort of agency to engage in all sorts of news with a significant degree of autonomy before the internet, this autonomy is very much impeded by Tüfekçi’s six components.

Not to talk about my internship over and over again, but while I was at the IPI, we had a super interesting panel on a topic very similar to Tüfekçi’s paper, titled, “Democratizing Algorithms for News in Fragile Democracies.” It was quite funny to hear the Human Rights consultant (?) at Google get attacked by social scientists and journalised based on the platform’s lack of action in democratizing algorithms as well as protecting journalists’ safety. Dropping the link here in case you’re interested.

Reading Response #3

Response to "The Ladies Vanish" and "The Truth About How Uber’s App Manages Drivers"

- What are some of the unique features of platform economies?

- Where should the responsibility of a platform stop with respect to the economic actitivities they enable?

Platform economies seem to have a uniquely unquestioned monopoly over the industries in which they operate. We seem to hold technological platforms significantly less responsible for many of the ethical consequences of their actions, for many reasons. Perhaps it’s because the complicated science of software and artificial intelligence, hidden as “black boxes” is way beyond many of us. Perhaps our lack of education on digital cutlure and the basics of computer science make us incapable of seeing that while a lot of platforms’ services really do work through individual machines and programmed softwares, there is still a huge amount of concrete human labour that goes into platform economies. The problem is just how well these underpaid and underrecognized workers are hidden.

For me, the question then becomes how can we shed greater light into the injustices platforms have caused (and will continue to cause if nothing changes) and place them on equal footing with any other monopoly overtaking a sector? We’ve seen it with Ticketmaster overtaking the online (and virtually offline) ticketing industry, and with social media giants monopolizing fake and/or politically misleading news for profit. It must become evident that not only should these black boxes become more transparent whether or not they are protected by millions of intellectual property laws, and also that these companies should be subject to just as many anti-trust laws as many others do.

Perhaps there should also be legislation addressing the hidden contract workers, described in “The Ladies Vanish”, encouraging a smaller base of permanent employees which may be subject to further protection than a large base of contract workers. The main point here is that the case for making algorithms transparent is imperative to keep these online giants just as accountable as their offline counterparts.

Reading Response #4

Response to "The Nature of Software" in Code/Space: Software and Everyday Life

- How is software different from other economic activities? How is it not?

- To what extent do you think that software affects its social/economical/cultural environment? Or the other way around?

I’ve mentioned this in almost all of my reading responses, but I’ll say it again, because my whole point is what we should be saying over and over again: software, AI, all of these big words, are just products made by humans that are hence inherently imperfect. I resonated a lot with one of the opening quotes to the article:

“The art of creating software continues to be a dark mystery, even to the experts. Never in history have we depended so completely on a product that so few know how to make well.” — Scott Rosenberg

The quote explains the dichotomy between the complexity and the humanity of the product while highlighting the confuddling devotion to it, with so few people really understanding how it works. In engaging with software, we must be aware of its man-made flaws and as we continue to depend on it in so many areas of our lives, we must increase awareness and knowledge about it. We cannot build the future of our society relying on technologies that only a supraminority of our population understands and can master. If we don’t make an effort to mobilize people to learn more about software and computer science we face the threat of an advancing technocracy through many topics we’ve talked about earlier through manipulative algorithm and platform governance.

Reading Response #5

Response to "Machine Bias: Risk Assessment in Criminal Sentencing" and "AI's Language Problem"

- What some of the things we should be concerned about regarding AI?

- What are some of the things we could be looking forward to regarding AI?

Both articles directly address some concerns regarding AI, while treating the solutions to the concerns as things to look forward to. I agree with this outlook. It’s important to recognize that the current use of AI in risk sentencing is quite dangerous and is implemented poorly: its racial bias has been proven. Yet this doesn’t mean that the benefits that it is believed to have should be completely scrapped, deleting AI from the justice scene completely.

One of our classmates at Sciences Po is an ardent unapologetic of AI completely replacing the justice system — perhaps you have heard his take. While I am completely against replacing the human legal force from its judges to its lawyers, I do believe that AI can and perhaps should be implemented in courtrooms both for efficiency and in order to eliminate humane-bias. But this also means that when AI will be used it needs to be used as just one of many primary, secondary, or tertiary sources used in the courtroom. We should treat the AI blackbox just as human-engineered as it is in practice instead of treating it as the all knowing God of rights and wrongs. We should use it in our considerations but never allow it to make any final decisions.

Why? The reasons are endless. The fact that no machine can ever learn mannerisms the way a human can, for example. Or no matter what dataset it takes to form its understanding of race, it will never become colour blind, and that it indeed shouldn’t. These are just some of the arguments for why it shouldn’t be the decisionmaker in the courtroom. But the language argument is just as potent for its other uses. AI can never be the superpower it is claimed to have the potential to be until it (And maybe we) understand language.

Particularly on this note of language I am particularly pessimistic. The problem is that words mean slightly different things to different people. Even within speakers of the same language, words’ lexical fields may overlap differently and overall have different meanings from one region to another. For many, it is the experiential associations with words that create preferences in the utilization of one word over another, creating even the slightest differences in meaning. But it’s not just these personal and experiential differences that make it so difficult for AI to interpret language. Some specific barriers are rooted in historically charged meanings for words based on identities. This ties us back to our earlier discussion on machine learning and the racial bias inherent in the American justice system, but one thing it also reminds me of is the very gendered aspects of many languages, and the way it differs completely from one language to another.

Many speakers of gendered languages cannot imagine a language where the gender of subject pronouns (or even objects or more depending on the language’s grammatical structure) is not specified. Further, upon realizing that there are in fact languages which do not have such a concept of gender in grammar, many assume that societies with such languages must be absolutely revolutionary. Perhaps a striking example of matriarchy, or “neutriarchy” (I just made that up). Yet my favorite example to give here is Turkish: Our language has no gender, yet we manage to use it to degrade women and undervalue their achievements over and over again. In fact, we do it in such a way that AI could never understand either: how can AI begin to explain these differences in languages even humans haven’t really studied or understood enough?

My answer isn’t original: I’m just repeating the idea that while there may be many potential benefits from AI they are yet to hit many man-made challenges as a result of the many biased results of our socialization we are yet to notice. Even the machines are made by people, and they are bound to influence their impacts at the end of the day.