Saturday, December 07, 2024

New Logics for Social Media and AI - "Whom Does It Serve?"

[UPDATED 12/17/24 to add Shaping the Future of Social Media with Middleware  (with Francis Fukuyama, Renée DiResta, Luke Hogg, Daphne Keller, and others, Foundation for American Innovation, Georgetown University McCourt School of Public Policy, and Stanford Cyber Policy Center (details below).] 

A collection of recent works present related aspects of new logics for the development of social media and AI - to faithfully serve individuals and society, and to protect democratic freedoms that are now in growing jeopardy. The core question is "Whom does it serve?"*

This applies to our technology -- first in social media, and now as we build out broader and more deeply impactful forms of AI. It is specifically relevant to our technology platforms, which now suffer from "enshittification" as they increasingly serve themselves at the expense of their users, advertisers, other business partners, and society at large. These works build to focus on how this all comes down to the interplay of individual choice (bottom-up) and social mediation of that choice (top-down, but legitimized from bottom-up). That dialectic shapes the dimension of "whom does it serve?"* for both social media and AI.

Consider the strong relationship between the “social” and “media” aspects of AI -- and how that ties to issues arising in problematic experience with social media platforms that are already large scale:

  • Social media increasingly include AI-derived content and AI-based algorithms, and conversely, human social media content and behaviors increasingly feed AI models
  • The issues of maintaining strong freedom of expression, as central to democratic freedoms in social media, translate to and shed light on similar issues in how AI can shape our understanding of the world – properly or improperly.

These works focus on how the 1) need for direct human agency applies to AI, 2) how that same need in social media requires deeper remediation than commonly considered, how 3) middleware interoperability for enabling user choice is increasingly being recognized as the technical foundation for this remediation, and how 3) freedom (in both natural and digital worlds) is not just a matter of freedom of expression, but of freedom of impression (choice of who to listen to). 

Without constant, win-win focus on this essential question of "whom does it serve?" as we develop social media and AI, we risk the dystopia of "Huxwell" (a blend of Huxley's Brave New World and Orwell's 1984).**  

  • New Perspectives on AI Agentiality and Democracy: "Whom Does It Serve?"
     (with co-author Richard Whitt, Tech Policy Press12/6/24) - Building toward optimal AI relationships and capabilities that serve individuals, society, and freedom requires new perspectives on the functional dimensions of AI agency and interoperability. Individuals should be able to just say "Have your AI call my AI." To do that, agents must develop in two dimensions:
    1. Agenticity, a measure of capability - what can it do?
    2. Agentiality, a measure of relationship - whom does it serve?
  • Three Pillars of Human Discourse (and How Social Media Middleware Can Support All Three) (Tech Policy Press10/24/24) - Overview of new framing that strengthens, broadens, and deepens the case for open middleware to address the dilemmas of governing discourse on social media. Human discourse is, and remains, a social process based on three essential pillars that must work together:
    1. Agency
    2. Mediation
    3. Reputation 
  • NEW: Shaping the Future of Social Media with Middleware (Foundation for American Innovation and Georgetown University McCourt School of Public Policy, 12/17/24) -- Major team effort with Francis Fukuyama, Renée DiResta, Luke Hogg, Daphne Keller, and many other notables, White paper building on this 4/30/24 Symposium that I helped organize, held at Stanford Cyber Policy Center. Assembled leading thinkers at the nexus of social media, middleware, and public policy. The only comprehensive white paper to offer a thoughtful assessment of middleware’s promise, progress, and issues since the 2020 Stanford Group paper. The goal is to operationalize the concept of middleware and provide a roadmap for innovators and policymakers. (The above two pieces extend this vision in broader and more forward-looking directions.)
  • New Logics for Governing Human Discourse in the Online Era (CIGI Freedom of Thought Project, 4/25/24- Leading into the above pieces, this policy brief pulls together and builds on ideas about how freedom of impression guides freedom of expression without restricting it, and how combining 1) user agency, 2) a restored role for our traditional social mediation ecosystem, and 3) systems of social trust all combine to synergize that process for the online era. It offers a proactive vision of how that can enable social media to become ever more powerful and beneficial "bicycles for our minds."
*Alluding to the Arthurian legend of the Holy Grail.
**Suggested by Jeff Einstein and teased in his video.
(revised 12/22/24)