Adobe Systems Pluralism and Disagreement 1 Automated Influence, Artificial Intelligence, and Democracy Jiří Baroš, Department of Political Science, Faculty of Social Studies, Masaryk University, Brno Adobe Systems Pluralism and Disagreement 2 Automated Influence, Artificial Intelligence, and Democracy Artificial Intelligence and Democracy What´s Wrong with Automated Influence? Adobe Systems Pluralism and Disagreement 3 Automated Influence, Artificial Intelligence, and Democracy Artificial Intelligence and Democracy What´s Wrong with Automated Influence? Adobe Systems Pluralism and Disagreement 4 Artificial Intelligence and Democracy AI and Dystopian Potential for Democracy Techno-Optimism AI and Public Spheres AI and Political Power AI and Economic Power Adobe Systems Pluralism and Disagreement 5 AI and Dystopian Potential for Democracy The Materiality of Human Affairs: Democracy vs. Autocracy Autocracies and Technological Advances Early Democracy and Advances in Production and Communication The Late 20th Century: Democracies Outperformed Dictatorships AI Offers Possibilities That Undermine the Conditions That Makes Democracy More Viable Than Autocracy (and Jacques Ellul) Adobe Systems Pluralism and Disagreement 6 Techno-Optimism Grand Democratic AI Utopia: The Large-Scale Involvement of AI in Our Collective-Choice Processes AI Reproduces Certain Features of Early Democracy While Solving Modern Democracy´s Distant-State and Overbearing-Executive Problems We Should Be Way of Letting Such Scenarios Guide Our Thinking (e.g., A Pure Intelligence?) Adobe Systems Pluralism and Disagreement 7 AI and Public Spheres What Are Public Spheres? Fourth Estate (Printing Press, Television, and…) Big Tech Companies (Wide Aperture, Low Deference etc.) Digital Media (and Artificial Unintelligence) Deepfakes Digital Public Infrastructure Adobe Systems Pluralism and Disagreement 8 AI and Political Power Profit-Seeking or Criminal Activities, Mass Protests vs. State Reactions (The Chinese Social Credit System, Scientific-Technological Elite, Intelligence, Racial Biases, etc.) How Digital Technologies Can Strengthen Democracy Preferences of Low- or Middle-Income Citizens… How Could Democracy Be Strengthened by Well-Designed AI Adobe Systems Pluralism and Disagreement 9 AI and Economic Power Egalitarian Empowerment vs. Economic Inequality A Worry That the Masses Would Expropriate the Elites, or the Opposite? AI Drives Inequality Optimistic vs. Pessimistic Sides of Predictions Surveillance Capitalism (Shoshana Zuboff) vs. Big Tech Companies Treated and Regulated As Public Utilities Adobe Systems Pluralism and Disagreement 10 Automated Influence, Artificial Intelligence, and Democracy Artificial Intelligence and Democracy What´s Wrong with Automated Influence? Adobe Systems Pluralism and Disagreement 11 What´s Wrong with Automated Influence? Big Tech – Automated Influence Two Perspectives Privacy Exploitation Manipulation Adobe Systems Pluralism and Disagreement 12 Big Tech – Automated Influence Big Tech Under Attack Automated Influence: The Use of AI to Collect, Integrate and Analyze People´s Data, and to Deliver Targeted Interventions Based on This Analysis, Intended to Shape Their Behavior. 2 Broad Forms: (1) Matching People with Products, Services and Content They May Find Appealing; (2) Tailoring the Message to the Individual Based on Their Inferred Susceptibility to That Method of Persuasion. Benign vs. Malign Face of Automated Influence Adobe Systems Pluralism and Disagreement 13 Two Perspectives The Interactional Approach: considers the interactions between agents that make up a social practice. It aims to identify adverse effects for individuals directly caused by those interactions. The Structural Approach: evaluates the emergent social structures of which those interactions are the leading edge. It considers how those social structures directly and indirectly impact people´s lives and their relational properties as well as their aggregate effects. The Interactional vs. Structural Critique of Social Practices and Automated Influence. Adobe Systems Pluralism and Disagreement 14 Privacy I Introduction: Training Data vs Targeting Data. Sensitivity and Data. (1) An Interactional Approach: Control of Date about You: You Do Not Have A Weighty Underived Claim to Unilaterally Control Your Intrinsically Nonsensitive Behavioral Data. Control Over Inferences: Identifying Patterns and Making Inferences from Licitly Acquired Data Is Not Itself Wrongful. Adobe Systems Pluralism and Disagreement 15 Privacy II The Role of Consent: The Inadequacy of Individual Consent to Legitimate the Collection and Use of Individuals´ Behavioral Data. The Only Meaningful Choice Is between Not Using the Internet at all and Submitting to Being Profiled and Targeted. Work Out a Collective Approach to Allocating and Using Behavioral Data => (2) A Structural Approach Adobe Systems Pluralism and Disagreement 16 Privacy III The Aggregate Effects of the Broader Institutions of Data Governance Surveillance and Sovereignty: AI Creates a Standing Incentive to Turn Everything into Behavioral Data. The Idea of Observation: to Build a Comprehensive Picture of Each Person. Ubiquitous Surveillance Makes for an Unfree Society. How to Ensure That Our Behavioral Data Is Genuinely Secure Against All Third Parties? Threatens Equality as well as Freedom: Some People Are Placed in an Extraordinarily Asymmetrical Position Relative to Others. Adobe Systems Pluralism and Disagreement 17 Exploitation I Exploitation Occurs When One Part to an Ostensibly Voluntary Agreement Intentionally Takes Advantage of a Relevant and Significant Asymmetry of Knowledge, Power, or Resources to Offer the Other Party Terms of Exchange to Which They Agree but Would Never Accept Were They More Symmetrically Situated in That Respect. Unfavorable Transactions: The Case That Individual Users Are Exploited by These Practices Rests on a Weak Foundation Predictive Models Depending on Massive Datasets Adobe Systems Pluralism and Disagreement 18 Exploitation II Dividing The Cooperative Surplus: The Cooperative Surplus Generated Through our Use of the New Digital Infrastructure Has Been Divided to Give Digital Service Providers a Disproportionate Share of the Benefit and the Power. Refusal and Resistance: The Systems of Automated Influence Worked to Atomize Individual Consumers… The Exploitation Objection Restated: As Self-Determining Political Communities We Do Have Robust Presumptive Rights to Set the Terms for how that Cooperative Surplus Is Distributed. Adobe Systems Pluralism and Disagreement 19 Manipulation I Manipulation Involves Undermining an Individual´s Decision-Making Power. The Wrong Manipulation Has Two Sides: (1) It Involves Effectively Subordinating the Will of Others (i.e., It Undermines Their Autonomy), (2) It Involves the Manipulator Placing Themselves Above the Manipulated, Treating the Manipulated as a Subordinate. Tailoring the Message, Targeting the Product How Effective Is Online Manipulation: Automated Influence Is Not Especially Effective on an Individual Level. It Can Be Significant in the Aggregate. Political Campaigning: Altering Subject´s Beliefs and Desires. Adobe Systems Pluralism and Disagreement 20 Manipulation II Stochastic Manipulation of Groups: The Magic and the Harm of Big Data Is in Its Aggregate Effects. In the Aggregate, the Interventions of AI May Make Nontrivial Impacts on Group Behavior as a Whole. It Preys on Some Pathologies of Collective Decision-Making: (Non)Political Decision-Making. It Concentrates Power in Too Few Hands and Pollutes Our Capacity for, and Willingness to Commit to, Collective Deliberation and Action. It Undermines Trust in the Authenticity of One´s Fellow Citizens´ Deliberations. Adobe Systems Pluralism and Disagreement 21 Manipulation III Democratic Deliberation and Collective Decision-Making: The Greater The Extent to Which Our Public Discourse Is Fragmented by Misinformation and Conspiracy Theories, the Less Capable We Are of Reasonable, Respectful, Collective Deliberation. Democratic Politics Becomes Nothing More Than a Thinly Veiled Struggle for Power. Epistemic Paternalism and Radical Centralization of Power. Adobe Systems Pluralism and Disagreement 22 Conclusion Developing Freely Self-Determining Political Communities So That We Can Be at Home in the Laws to Which We Are Subject. The Mixtures of Chaos and Untrammeled Power Has Led to Seriously Deleterious Effects (…). The Problems Are Collective Action Problems, the Presumptive Solution: More Power, Not Less. Impasse: Leaving Power with the Titans of Tech vs. Giving the Same Kind of Access to National Governments.