My notes for a session on "***Cybersecurity, privacy & safer Internet***" at [NetMission Academy 2022](https://netmission.asia/2021/12/13/netmission-academy-2022-result-announcement/)
## Content
#### Who is Prateek Waghre?
- Researcher at Takshashila.
- Working on Digital Communication Networks
- Public Policy misfit - got here after spending nearly a decade doing Technical Consulting and Product Management
- Most work has been on Internet Shutdowns / Disruptions and Information Disorder / Ecosystem.
- Also writes MisDisMal-Information - publication on the information ecosystem in India (mainly) : [techpolicy.substack.com](https://techpolicy.substack.com)
#### Base Framework
Adapted Balkin's Conceptualisation
![[Balance of Power TIE.png]]
- The adversarial space construct
- Once you define/think of a solution that may address the specific harm you are concerned out.
- List out the ways it can be used to exacerbate that harm.
- Then, zoom out, and list out the ways in which it can make other harms worse.
#### What are Digital Communication Networks (DCNs)?
Composite entities consisting of
- Capability: Internet-based products/services that enable instantaneous low-cost or free communication across geographic, social, and cultural boundaries. This communication may be private (1:1), limited (1:n e.g. messaging groups), or broad (Twitter feeds, Facebook pages, YouTube videos, live streaming ), and so on.
- Operator(s): Firms/groups that design/operate these products and services.
- Networks: The entities/groups/individuals that adopt/use these products and services, and their interactions with each other. Interactions can be on the same service, another service or on no service at all (offline).
Two defining aspects of DCNs are *low entry costs* and *reliance on user-generated content*
[Working Paper Link](https://takshashila.org.in/takshashila-working-paper-governance-of-dcns-i-categorisation-of-harms/)
###### Why this approach?
- ICTs, social media networks, social media platforms, Big Tech, Digital Platforms + many more
- Context-dependent
- Too broad, too narrow
- This framing should encourage their study from the perspective of their effects on societies as a whole, rather than focusing on specific companies, technologies, sharing mechanisms, user dynamics, and so on— attributes which are constantly evolving.
- Recognising underlying dynamics such as *'Misinformation is networked, content moderation is not' - Renee DiResta*
- Researchers have called for changes in Global Collective Behaviour to be studied as a crisis discipline
- Rapid changes in the scale and structure of human networks
- Impact on the quality of information in circulation
- The role played by algorithms in directing the flow of information
- Focus on the effects of DCNs in the spheres of competition in markets, data gathering, and narratives in public discourse.
![[Pasted image 20210727165957.png]] ![[Pasted image 20210727170008.png]]
Exclusions and gray zones. Some examples
- Gray Zones
- Working conditions of content moderation staff
- Information deficits on google search
- Exclusions
- Competition in e-commerce marketplaces/service aggregators, and Labour conditions for gig-economy workers
#### Harms
###### Why classify and/or categorise harms?
- Since 2010, internet pop increased from 30 - 66%. In 2021, 4.2B social media users estimated to spend 3.7T hours. Yet, ~half of the world's population is on.
- In India, even with 795M, 59 connections per 100 people. Headroom for adoption still exists. **scope for more pronounced effects of the harms**
- Similar dynamics of global information flows impact different societies in different ways.
- Thus, if the goal is to limit their harms, it would be helpful to try and collate and contextualise them, to whatever extent that is possible.
##### Classification
![[Pasted image 20210727172722.png]]
###### Competition Harms
- Stifling innovation and competition
- Track/collect information on competitors
- Withhold data
- Funding for newer entrants
- Dictating terms of business
- Power imbalance between firms that operate DCNs and businesses that rely on them
- Forced compliance to business terms
- Unequal data access, preference to certain partners
- Rivalling platform participants
- Self-preferencing
###### Data-related Harms
- Privacy
- Consumers likely to accept policies that erode privacy given reliance on DCNs
- Have limited clarity/visibility on how data will be used
- Honeypot, not immune to breaches / leaks
- Individual agency
- Limited choice when it comes to accepting EULAs/user-agreements
- Opt-out limited by choices provided
- Dark patterns / behavioural nudges
- Algorithmically curated content - limited choice to accept/reject content
- Attention economy
- Data value chain
- Firms operating DCNs are concentrated in US (and now China). Storing and processing domestic data in other countries raises questions about its ownership and about user privacy, and has implications on national security
###### Narrative Harms
- Actions of firms operating DCNs v/s scaling effect
- chilling effect on freedom of expression through content takedowns or inaction, resulting in the silencing of vulnerable groups
- regulatory capture through political lobbying
- boosting or stifling particular viewpoints as a result of design choices, with the potential to exercise power over content dissemination
- virality of mis/disinformation
- coordinated attacks and inauthentic behaviour
- proliferation of CSAM and terrorist propaganda material
- Incentives
- Attention economy incentivises grabbing attention. Firms focus on maximising time spent and boosting engagmenet metrics.
- social-validation feedback loop ( desires for connection, competition, FOMO-avoidance and habit creation), and facilitating digital behaviour contagion. Intersection if data-related harms.
- Tendency among participants to ‘perform’ – sensationalism and tribalism are rewarded with engagement. Conversely, penalty for not participating.
- Benefits accrue to participants near the top. Esp. in politics. - celebritisation of politics.
- Amplification of harmful messages and behaviours.
- Control and Manipulation of Information
- content moderation + lack of transparency can impact/supress legitimate voices
- Cooperation/Kowtowing/inadvertently providing information to states can lead to supression of dissent
- coordinated harassment, targeting,trolling, swarming, or doxxing by financially, ideologically, or politically motivated actors esp against vulnerable groups may stifle viewpoints so that they are no longer expressed publicly.
- Political weaponisation of mis/disinformation, distraction, information overload have become easier.
- Intersection with data as some of them rely on tracking and targeting / microtargeting individuals and groups + newsfeed DCNs.
- Trust and Integrity
- Information Integrity
- News Media - shift in consumption model, atomisation, competiting with lower quality information, reduced ability to monetise, dependence on advertisers, resort to partisan positions/coverage, speed over accuracy.
- Institutional trust - democratic, scientific institutions
- effects can be both short-term and long-term, and they can be particularly abused by domestic and international adversaries to subvert the integrity of civic or democratic processes.
- Social Cohesion
- Homophily, clustering, hyper-personalised information diets, human biases and context collapse contribute to lower levels of empathy, increased tribalism, polarisation, radicalisation of discourse - mis/disinformation, incivility, hate speech
- Polarisation : attitudinal (consideration of evidence), affective (disliking individuals/groups that hold different/contrasting views) and knowledge (disagreement on shared realities)
- Escalatory / competitive cycles of hate speech, fear-mongering, spread of conspiracy theories, incitement, physical violence
##### Categorisation
| Market Failure | Social Problems | Cognitive Biases |
| :----------------------------------------------------------: | :----------------------------------------------------------: | :----------------------------------------------------------: |
| market power<br />market power - gatekeeper<br />market power - leveraging <br />market power - information exploitation<br />information asymmetry<br />externalities | addiction<br /> child abuse<br />civic integrity<br />crime<br />harassment<br />sexual abuse/harassment<br/>inequity<br />information integrity<br />misaligned incentives<br />privacy impact<br />radicalisation<br />social division<br />suppression | anchoring bias<br />availability cascade<br/>bandwagon effect<br/>blind-spot bias<br />confirmation bias<br />conservatism bias<br/>information bias<br/>recency<br />stereotyping |
###### **Market Failures**
![[Pasted image 20210730080443.png]]
![[Pasted image 20210730090502.png]]
###### **Social Problems and Cognitive Biases**
Social problems: in the context of “less than ideal situations” affecting many individuals/groups within a society, which need to be addressed through “collective effort or action”. While specific problems such as crime, poverty, and sexual abuse are recognised as social problems, many of the harms we identified do not fall neatly into these categories.
![[Pasted image 20210727184930.png]]
![[Pasted image 20210727184947.png]]
![[Pasted image 20210727184959.png]]
![[Pasted image 20210727185018.png]]
## Related
1. [Takshashila Working Paper – Governance of DCNs I: Categorisation of Harms – The Takshashila Institution](https://takshashila.org.in/takshashila-working-paper-governance-of-dcns-i-categorisation-of-harms/)
2. [Takshashila Discussion Document – Governance of DCNs II: Opportunities and Benefits – The Takshashila Institution](https://takshashila.org.in/takshashila-discussion-document-governance-of-digital-communication-networks-ii-opportunities-and-benefits/)
3. ["Privacy Harms" by Daniel J. Solove and Danielle Keats Citron (gwu.edu)](https://scholarship.law.gwu.edu/faculty_publications/1534/)
## Colophon
%%
title:: NetMission Academy 2022 – Training 4- Cybersecurity, privacy & safer Internet 2022-01-27
type:: [[output]]
tags::
url::
file::
creator:: Prateek Waghre
%%
created:: 2022-01-27
status:: [[brewing]]