Undermining Democracy Essay
How will AI affect politics in the future?
(social credit scores involve big data/ai analytics. Financial Credit score), AI – loans. Explainability/Transparency. Facial recognition.) Vs Right to reputation, privacy, explainability
Automation – unemployment
Emotions can undermine logical or rational considerations in some circumstances – Utility of AI.
Undermining democracy
Theme: How will AI affect politics in the future.
Technology and Politics – misinformation, digital bias
We are assessing the impact of technology shifts on politics.
AI’s societal impact, how technology is influencing political governance today, and how these changes are shaping our lives – for better or worse.
Mistrust in human intuition. Why do we make poor decisions: Ego depletion, The halo effect, Belief overkill, the law of small numbers.
This phenomena can be addressed with AI data analysis but actors in the social media space often take advantage of these human flaws to gain attention. The briefness and volume of content and the way we tend to engange with it (very quickly) arguably encourages some of the phenomena above. Lack of critical engagement with topics, polarised content, irritable, faceless and highly responsive audiences.
AI futures framing – AI revolution/ the age of AI. Presented and framed by governments (Germany/China), EU UN have drafted their AI strategies.
Conc
Does this technological supremacy go deeper into our realities than we think. What room is there for debate?. Which version of reality is correct. To what extent can we create the reality we live in. How does our interactions with technology dictate our perception of reality and perceptions of others. How powerful is open source information and how long does it take to become publicly available.
Alpha go – microcosm for effective learning and incredibly powerful descsions/outcomes
AI Technologies
AI driven text generation (NLP), Social Media chatbots both create artificial personas (what problems for politics does this pose?)
- Pink Slime Journalism.
- Comments/ Social media posts – fake tweets (Brexit/Trump Paul Mercer)/drive polarisation.
- Eg – NET neutrality comments.
- Paul Mercer Renaissance Technologies funded Epoch 900,000.
- The epoch times spread disinformation during the trump campaign
What is ai – it is a hyped term, 1956 dartmouth conference
Royal soc talk suggests AI is 3 things – technical approaches, social practices, industrial infrastructure.
We have gone from symbolic logic to ML. ML is a constellation of technical approaches from statistics and gradient based optimisation.
Political Issues
Employment vs Automation
Loss of values/traditon/morality Reactionary responses from left and right wing – mainly populists
BIAS
Economic security – integrity of financial markets
National Security – what is the capacity of nonstate actors to penetrate systems/ undermine political systems.
Inequality
Ability to monitor media with articifal intelligence – campaigns – faster, better decisions
- Officaials
The transfer from academia to technology is process that requires scientific breakthroughs
RNC – Obama – ability to monitor campaigns highly increased.
AI Can:
Improve operational efficiency. Governments require an vast and increasing levels of data processing. The tax division and Exec office of US attourneys uses Veritone’s AI tools.
Automate fact checking
Helps candidates monitor opponents
Helps politicans communicate with their constituencies.
Veritone Politics – one of the companies that turned big data into actionable intelligence. The ability to monitor media at scale helped them alter outcomes of campaigns that they were managing.
Targeting votes online
Other than campaigns the ability to monitor home states/consitiuencies/ gives them more optics that lets politicians be more efficient in their jobs.
How available is AI to smaller less well funded parties?
The Russia Effect –
The Facebook Effect
The biggest bias we face in society
Attention based models
Scoop’s contention – Untamed AI will continue to drive polarisation and division.
Reccomendation algorithms reduce free speech. Without censoring content we can restrict the mechanisms by which these algorithms are trained which create echo chambers, polarisation, psycological manipulation, forgoe ethical considerations
The greatest threat to our political future coming from the field of AI is The combination of attention and recommendation based algorithms poses the gre
Sean Parker – ‘How can we consume as much of your time and concious attention as much as possible’ (persuasive design model, hook model)
Stanfords persuasive technology app – the original idea was to promote good behaviours like quitting smoking. \
^this is an unethical design
Operant conditioning. The psychology of persuasion is being applied to technology. A positive intermittent reinforcement, creates unconscious habits– operates like a slot machine.
The server farms
SD – notes
People are the product, their attention is the product sold via ads.
There is a misconception that they sell our data – they use this data to create models & predictions. Which emotions trigger you, what content will keep your attention for the longest period in time.
Attention goal – keep you on
Social goal – increase number of users
Ad goal – generate revenue
Internet companies are the richest in the history of humanity
Online connections has becom primary – we have a global generation of people who are raised within a context where the very meaning culture is manipulation.
There is an entire discipline called growth hacking. Facebooks growth tactics have become the standard for Silicone Valley.
Demanding, Seducing, Manipulating.
Key processing power expanding exponentially – get evidence for this.
Algorithms are optimised to some definition of success.
When technology overwhelms human weaknesses – addiction, polarisation, radicalisation, outrage-isation, vantity.
Most of us accept the reality that is presented to us.
Disinformation – we have a system that biases towards false information. False information spreads faster makes more money.
very few laws on digital privacy. Plenty of laws for phone numbers.
Corps using powerful AI to hold our attention. It’s the greatest distraction.
Sources:
1
2
MCEVOY, F.J., 2019. Political Machines: Ethical Governance in the Age of AI. Moral Philosophy and Politics, 6(2), pp. 337-356.
Epistocracy – (Brennan) – a system that priviliges the most politically informed citzens’ votes.
Principle of competence requires governments to be competent with their political decision making – AI is being adopted for exactly this reason. To improve Efficiency and efficacy is a moral obligation. Shu, Tsay, and Bazerman (2012, pp. 243ff.) similarly present evidence to counter the myth that policymakers are rational actors. They aggregated research to show that ‘when considering issues of high complexity, decision makers are typically constrained by time, imperfect knowledge, and overreliance on general rules of thumb.’ Their supporting argument cites, among other factors, the prevalence of loss aversion which supports a tendency to focus on losses rather than gains when contemplating change. This ultimately translates into a greater concern with the risk of change than the risk of failing to change. Shu et al. say this status quo bias can lead to a dysfunctional desire to maintain a broken system. There is a great need to look at issues in the best analytical way possible. Lastly, there are also a number of more conscious factors that can derail the good decision-making of political figures and policy experts. Namely, their concern with acceptability and public approval which comes to distinguish them from analytical decision-takers whose behavior is driven by the need to maximize utility in general (Farnham 1990, p. 99).
3
4.
5
6
7. https://datasociety.net/pubs/oh/DataAndSociety_MediaManipulationAndDisinformationOnline.pdf
8. https://scholar.harvard.edu/files/shoag/files/cell_phones_and_motor_vehicle_fatalities.pdf
Shows links between social media and declining quality of life. ??? really bitch
9. TWENGE, J.M., HAIDT, J., JOINER, T.E. and KEITH, C.W., 2020. Underestimating digital media harm. Nature Human Behaviour, 4(4), pp. 346-348.
Shows ‘heavy social media use is consistently associated with negative mental health outcomes.
10. http://www.diva-portal.org/smash/get/diva2:1176971/FULLTEXT02
A Quantative Study of Social Media Echo Chambers
Joakim Johansson
Claims that require Evidence
- The rate of AI capability is increasing – someones law
- The amount we interact with AI suggested content is increasing
- The amount number of echo chambers is increasing
- Polarization is growing and can be attributed to alorithms
- Where/How do people consume their news get stats on search engines, news outlets, social media. This should show a large and increasing dependence on recommended news.
‘a large—and ever-increasing—share of news consumers rely on algorithmically curated environments in which algorithms automatically select personalized news based on information about individual news consumers.
- Theorisation of echo-chambers vs epistemic bubbles
Furthermore, echo chambers can explain what epistemic bubbles cannot: the apparent resistance to clear evidence found in some groups, such as climate change deniers and anti-vaccination groups.
Many users do not know about the exist- ence of algorithmic personal ltering. Even amongst those that do, most do not have access to the particularities of the algorithms doing the ltering; thus, the very opacity of the process makes it harder for a user to successfully evaluate and epistemically com- pensate for such ltering (Miller and Record 2013). Thus, most users signicantly under- estimate the degree to which their exposure to information, via search results, has already been tailored to present information to which the user will already be amenable.
They offer the following denition of an echo chamber: an echo chamber is a bounded and enclosed group that magnies the internal voices and insulates them from rebuttal (Jamieson and Cappella 2008: 76).
Much of the recent research on causes of belief polarization focuses on the causal role of individual psychology, such as the tendency towards laziness in the scrutiny of one’s own beliefs (Trouche et al. 2016). Similarly, recent studies on climate change denial focus on studying the relationship between an individual’s stated political beliefs and their reactions to climate change information, without inquiring into the social epistemic structures in which the individuals are embedded (Corner et al. 2012)
One of the most important questions about AI is what it will do for politics. Some have argued that AI could be used to make political decisions more objective, but there are many other possibilities. For example, some argue that an algorithm can replace a politician’s judgement and help them decide how to vote on issues such as climate change or immigration. However, this would not lead to a system where humans are entirely replaced by machines – rather they would just be co-opted into making these decisions in ways that benefit their own interests (such as when algorithms recommend companies hire people from certain backgrounds).
Theme
Theme: How will AI affect politics in the future.
Technology and Politics.
1-3 Echo-chambers, big data and protest Papers: Alexandra Cotofana (Butler University)
#REZIST. Romanian Digital Protests and Lessons in Imagined Liberalism.
Abstract: In January and February of 2017, Romania experienced some of the most numerous online and street manifestations since 1989. Close to half a million people in Bucharest and more in other cities across the country occupied the streets, protesting against a bill that the government was attempting to pass. The bill had the potential to decriminalize acts of corruption committed by state administrators. While there is much to be applauded about the Romanian protests that took place at the beginning of 2017, the media did not elaborate on many of the discursive aspects, and failed to explain the historical milieu, otherwise vital to understanding the complexity of the protests. The national press mostly focused on taking sides, while the international press focused on highlighting the starch difference between the corrupt state apparatus on the one hand, and the citizens’ intent on enforcing the rule of law, on the other hand. Yet this black and white portrayal leaves much
1 At the moment I write this paper, the preliminary results of the elections are only known, but opinion polls predict Zelensky’s victory.
4
to be desired for the scholarly community at large, and cannot explain the presence of fascist, anti-Semitic, and racist slogans chanted by some of the protesters.
The talk examines the effects of right-wing digital militarism and discriminating data by discussing recent events in Romania and elsewhere. The piece questions the concepts of ‘protest’ and ‘civil society’, that are internationally acclaimed as beacons of democracy, together with their digital lives. Despite online spaces, protests, and the civil society all being associated with the imagined singular progress of modernity, I argue that their intersection creates a digital subterfuge for discrimination algorithms and the proliferation of fake news. Furthermore, topically, the critical analysis of discourses and politics of the occult creates alternative data literacies, that can help us understand the ways in which online echo-chambers are proliferated.
Hande Eslen-Ziya (University of Stavanger)
From anger to solidarity: The emotional echo-chamber of Gezi Park Protests.
Abstract: By using Gezi park protests that took place in Turkey in 2013 as a case study the paper explores the significance of echo-chamber of emotions involved in protests. The analyses will be composed of thirty-seven interviews conducted with activists on how they define and understand their preferences and motivations for protesting. The paper will discuss how participants get influenced by beliefs, motives and opinions and will bring forth the changes of emotions during such process. By employing social constructionist approaches to emotions, the paper will discuss three socialization processes: (1) views drawing on the emotional template of Gezi Spirit; (2) belonging – forming intense affective ties within the social movement in action; and (3) re- conciliation—staging expressive public rituals of reconciliation between groups that previously had opposite interests and help create an emotional echo-chamber.
Derya Gül Ünlü (Istanbul University)
How Do We Express Our Digital Emotions?: Reviewing Emoji Use Trends from Gender Perspective.
Abstract: With the popularization of digital communication tools, individuals have started to take part in digital media more, which increased their involvement in various visual and written digital communication process. The chance individuals obtain to send visual and written messages to each other in a fast way brought along changes in the ways conventional written language is used. Today, individuals can convey their ideas or mood to the other party by sending various visual messages easily. Emojis are the most prominent tools among these visual messages. Emojis represent an important cultural development that takes place in language towards a more visual communication, and provide an opportunity to convey messages in a more creative way (Şener & Atar, 2017; 199). On the other hand, many researches reveal the fact that there is a gender based differentiation between women and men’s manner of speaking according to gender role expectations (Cameron, 1994; Fishman, 1978; Lakoff; 1973; Spender, 1980; West, 1984). Considered in this context, it is important to determine whether there is also a gender based differentiation between the ways women and men use emojis. With this purpose, the research aims to determine whether there is a difference between women and men in the ways they express their digital emotions according to gender expectations by means of a questionnaire based field research. It is considered that the study will be important in terms of addressing individuals’ emoji use tendencies from a feminist perspective.
Structure
Title How will AI affect politics in the future?
What are the foundations of AI? (Explain what AI is in simple terms, discuss its ascendancy through media,data,cloud,tech, )
AI is a profoundly concentrated industrial infrastructure.
Is ai a political tool.
What are the ways in which AI has affected politics in the past?
Campaigns: Brexit, Trump – AI analytics.
Media – how we engage with it/ targeted ads/media/ seems to fuel polarisation.
MNC’s/Tech Giants – are at the forefront of AI
Bias in policing
Authoritarian regimes like China, Russia
What are the trends in AI that seem most likely to contribute to the political future?
One of the most important questions about AI is what it will do for politics. Some have argued that AI could be used to make political decisions more objective, but there are many other possibilities. For example, some argue that an algorithm can replace a politician’s judgement and help them decide how to vote on issues such as climate change or immigration. However, this would not lead to a system where humans are entirely replaced by machines – rather they would just be co-opted into making these decisions in ways that benefit their own interests (such as when algorithms recommend companies hire people from certain backgrounds).
