Social Media Harms and Solutions

Overview

Panel discussion about the documentary "The Social Dilemma" and social media harms.
Speakers include filmmaker Jeff Orlowski and experts: Tristan Harris, Tim Kendall, Cathy O'Neil, Rashida Richardson.
Focus: how platforms shape behavior, addiction, algorithmic bias, polarization, misinformation, and possible solutions.

Technology companies optimize for attention and engagement, not social good.
If a service is free, advertisers pay and users' attention/value is monetized.
Design techniques exploit human psychology (Pavlovian responses, lizard brain).
Fake news and misinformation spread faster than truth, amplifying polarization.
Recommendation systems create individualized “Truman Show” realities, destroying shared consensus needed for democracy.
Algorithms embed and propagate historical biases, leading to unequal outcomes (loans, hiring, insurance).
Global expansion without local infrastructure or moderation leads to grave harms (example: Myanmar and Rohingya).
Companies often lack diverse perspectives and stakeholder input during design and deployment.
AI and algorithmic moderation are not silver bullets; they can be slow, subjective, and insufficient.
Short-term product fixes are band-aids; long-term structural change is required.

Photo tagging on Facebook increased engagement by triggering social self-consciousness.
Ad tech targeted predatory ads (for-profit colleges aimed at single poor Black mothers).
Mortgage risk models and AAA mortgage-backed security ratings were algorithmic failures.
Myanmar: Facebook-built infrastructure amplified government propaganda against Rohingya.
Political microtargeting: campaigns know more about users than users know about campaigns.
TikTok recommendation engine can amplify or suppress content clusters (e.g., anti-vax content).

Attention Economy: Business model where user attention is the main commodity for monetization.
Filter Bubble: Personalized information environment that reinforces existing beliefs.
Recommendation Engine: Algorithmic system that selects and ranks content to maximize engagement.
Algorithmic Auditing: Process to evaluate algorithms for bias, fairness, and compliance.
Predictive Policing: Use of historical data to predict future crime locations or persons.
Digital Colonialism: When platforms dominate internet access and content in countries lacking local infrastructure.

Individual: addiction, anxiety, depression, deterioration of youth mental health.
Social: polarization, erosion of trust in expertise and institutions, decreased shared reality.
Economic: unequal access to opportunities (jobs, mortgages, insurance) via algorithmic segregation.
Political: targeted misinformation, voter suppression, foreign interference, weakening democracy.
Global: platforms displacing local media ecosystems and enabling harmful government actions.

Individual and community actions: watch the film with politically different people and compare feeds to build empathy.
Platform product fixes (examples): turn off algorithmic amplification, remove trending topics, de-emphasize misinformation.
Transparency: demand disclosure about algorithms and their effects; require platforms to show how recommendations work.
Small-scale policy: algorithmic audits for hiring platforms, mortgage systems, and other economic decision systems.
Cultural movement: group migration and collective actions in schools, communities to shift norms.

Rethink business model: move away from attention-maximizing ad models toward models aligned with public good.
Regulation: create rules or an oversight body (analogous to an FDA for algorithms) to force platforms to follow public-interest laws.
Multi-disciplinary design: require stakeholder inclusion and diverse viewpoints before algorithm deployment.
Sectoral approaches: combine civil rights enforcement, antitrust, privacy protections, and new algorithmic governance.
Global coordination: address transnational scale of platforms and foreign influence on information ecosystems.

Watch "The Social Dilemma" to understand concepts and share with people who disagree politically.
Learn basics of how recommendation algorithms work and how they can be manipulated.
Practice digital hygiene: monitor screen time, experiment with notification settings, try collective device-free periods at school/home.
Support calls for transparency and algorithmic audits at local institutions (schools, employers).
Study regulatory proposals and follow news about platform policy changes and enforcement (privacy, antitrust, civil rights).

The current attention-driven model fuels addiction, polarization, bias, and political manipulation.
Algorithms are not neutral; they reflect decisions, values, and historical bias.
Small, implementable fixes exist, but systemic change (business model and regulation) is required.
Collective cultural action plus policy intervention offers the best path to restore shared reality and reduce harms.