How to Think About Making AI and Social Media Pro-Human

The recent US midterm elections delivered a mandate to change course on AI. Voters elected new leaders for federal and state offices that they believe will favor people over Big Tech and take a more muscular approach to regulating AI and data centers. The popular demand for intervention is now unmistakable. But what kind of intervention is necessary?

New alliances are popping up to try to answer that question. For instance, a new “Pro-Human AI Coalition” of more than 60 organizations launched just weeks before the midterms. It joins much good work already underway to return technology toward the path of augmentation — using technology to make people more capable, rather than replacing them.

But first in social media and now in AI, those efforts have mostly been siloed, fragmented, little-recognized, and under-resourced. I have been arguing for an alliance like this for some time, after seeking technology that augments human intellect rather than replacing it since 1969.

Now there is new hope for realignment, spurred by widespread fear that we are at a tipping point. For those who want technology to be pro-human, it is time to treat the new problems of AI and the still unresolved problems of social media as related parts of the same challenge: building a digital ecosystem that serves people and democracy, not corporate platforms.

This is one of the great challenges of our age. Over the past decade of social media, as it became clear that the technology was weakening human thinking and collective decision-making rather than augmenting it, many have come to understand that this is not just a technology problem, but a broader sociotechnical one. There are many technical and nontechnical levers to reverse it, but no silver bullet. I see three overarching disconnects that a coalition could help fix — and offer a few simple framing principles to address them.

Disconnect 1: AI and social media are merging

The first disconnect is that AI is not separate from social media, but is fusing with and absorbing it. The so-called “AI” and “social media” fields must work together to build an ecosystem that serves people. AI already shapes what we see on social media, it is starting to take over some of social media’s role in filtering information, and it draws on social media content (which is very uneven in quality) for training. So far, the hyperscaling of AI has largely overlooked the fact that humans evolved and flourished as social creatures — and to the related lessons of social media’s failure to serve real human connection and institutions. At the same time, the platforms we call “social media” are becoming increasingly less social, suggesting a need to face a new reality of "post-social media” or “parasocial media.”

But can humanity survive and flourish in a “post-social” world of people talking to computers, rather than one of people talking to people, augmented by computers? To channel RFK Sr., who borrowed from George Bernard Shaw, media scholars must help us see things as they are, and ask why, and media architects must dream of things that can be better and ask, why not? For AI to become pro-human it must be not only pro-individual but also pro-community — it must help people connect with each other. Think of this as “AI/SM” (AI and social media as one), or as “intelligent, pro-social media” (using “media” in McLuhan’s broad sense).

I hope the Pro-Human AI Coalition will also address the inseparable cause of pro-social media. That would seem consistent with the April report from Project Liberty (a founder of the coalition) and the Social Web Foundation (PL-SWF), which makes a clear case for that holistic focus, surveying key issues raised as AI and social media merge. Whatever one’s assessment of existential risk, societal risk is already a clear and present danger common to both AI and social media.

Disconnect 2: Tech is ignoring our social nature

The second disconnect is that neither mainstream AI nor social media are being built to support the rich social dynamics essential to humanity — the nuanced interplay between individuals and the dense network of communities and institutions that has co-evolved over millennia. Some scholars propose addressing this by combining siloed technologies that can reinforce each other (notably “AI and the future of digital public squares,” by Beth Goldberg and 26 coauthors; “Prosocial Media,” by Glen Weyl and others; and “Improving social media for democratic discourse,” by Fan Cheng and others). Some civil society organizations take a similarly broad view, again notably including the PL-SWF report.

Disconnect 3: Remediation requires a whole-of-society project

A third disconnect is that even as these ideas gain recognition, putting fixes into practice is difficult. There are an overwhelming number of them, the devil is in the many details, and broad-spectrum reform is a complex sociotechnical process gated at human speed.

I have previously proposed a broad, pro-human AI and social media coalition to close these disconnects: “a loosely federated, ‘big-tent’ alliance for a portfolio approach to motivate and guide the ongoing whole-of-society project needed to evolve a healthy sociotechnical ecosystem.” The newly formed Pro-Human AI Coalition promises to be the start of that — and can help guide a new awakening in US policy.

My suggested high-level architecture includes emerging support tools for social mediation processes (how others help us filter and make sense of information), individual and community control of attention, community representation, and a relentless focus on whose interests AI serves. This architecture offers a positive vision that goes beyond fixing problems. It would empower people and the collective intelligence of their many overlapping communities to select, govern, build, interconnect, and evolve the agents and services that shape how we think together, so that they serve everyone.

This Part 1 provides an overview of a few guiding principles. Part 2 expands on how they function to support pro-human objectives and then concludes with a call for a “big tent” alliance to obtain agreement on — and guide ongoing development toward — those shared principles. A supplement provides additional notes, references, and updates.

Focusing on a few big-tent framing principles

These are wicked problems with no simple solution. They are far too complex, nuanced, and context-dependent for any one person or group to fully understand. Additionally, progress has been slow because dominant platforms’ business models and centralized designs underweight the public good and user rights, while policy responses have remained limited. What can shape and hold a diverse portfolio of efforts together is a small set of shared principles that work at different levels: thought as a social process, three pillars of healthy discourse, negotiated access to user attention and data, loyal personal (and community) AI agents and a simple, essential and overarching question: who does it serve? These principles are aligned with the more detailed Pro-Human AI Declaration, but meant to be simpler and more universally compelling.

Principle 1: Thought as a social process

Debates rage over expression as a matter of “free speech rights” in isolation, often on a “for me but not for thee” basis. But the real issue is the breakdown of this broader process of how social creatures think. Online media harms are commonly viewed as expression management problems, but it is managing attention where we have lost our way. Attention had been organically manageable because social influences helped guide it. Now, faced with more content creation than anyone can process, we need feed algorithms, or “attention agents,” to help us. Platforms have taken over that role, but their algorithms fail to guide attention in the nuanced way human social mediation once did.

Historically, word of mouth and other forms of social mediation naturally helped us self-regulate both expression and attention. But now platforms control feeds and reputation signals such as likes and shares. Their engagement-optimized systems redirect our social feedback systems toward maximizing engagement, which often diverges from what users and communities need, while marginalizing important components of our social mediation ecosystem.

Principle 2: The three pillars of online discourse, linked by open interoperability

The coalition calls for “Designing a Pro-Human Al Stack…Building architecture for Al systems that are safe, open, and fair.” I suggest the Three Pillars illustrated in Figure 2 as important, mutually reinforcing aspects of that stack. They show how healthy discourse works: individual and social processes that guide each other, plus the reputation that emerges from and feeds back into them.

The first pillar is about individual choice and agency: the primal human process that steers our inputs through our attention, drives individual thought, and outputs expressed ideas to the network. The second is the social mediation ecosystem: the deliberative (or reactive) feedback processes by which others, both human and machine, mediate message contexts and flows from speakers to listeners, much as in the earliest word-of-mouth propagation. The third is reputation and trust: distillations from the other pillars, which feed back to them, evaluating whether speakers and mediators’ context are worthy of attention (relevant, valid, and desired).

All three pillars are essential, because each one has weaknesses the others can balance. Individual choice and agency is the focus of much reform effort, but there are valid concerns that, if poorly-mediated, it can lead to echo-chambers, polarization, and extremism. Social mediation is how society has always “moderated” (not eliminated) individual excesses, but without sufficient individual agency, it can drive groupthink and tyranny of the majority. Reputation and trust connects the two pillars in a feedback loop, distilling what people choose, share, and comment on into a kind of socially-derived metadata that guides further attention and interaction.

Open interoperability, which enables what is often called middleware (software that sits between users and platforms), is the foundation for that interplay. As code increasingly shapes discourse and cultural evolution, interoperability must be built in to preserve the agency, flexibility, extensibility, and resilience that humans and our social systems historically developed more organically. Monolithic, centralized walled gardens are too rigid for that, no matter how benevolently intended.

Beneath this foundation sit the networking platforms themselves, along with a layer of technical foundations such as alignment, integrity, transparency, privacy and security, resourcing foundations such as revenue and incentives models, and civic foundations such as regulation, standards, and civics education — all of which are important to being pro-human.

Principle 3: Negotiated access to user attention and data

Much of the harm of online media and the failure to seriously address it arguably stems from the extractive ad-based revenue models that dominate current social media — and may soon dominate consumer AI services. The Internet's “original sin” is not advertising itself, but abusively extracting user attention and data with little or no user control — or compensation. Alternative funding models are also desirable, but reforming the ad model, which currently incentivizes harmful engagement over quality might be one of most expeditious solutions. We should aggressively pursue both.

Users should control the terms by which others can access their attention and personal data, including preferences and intentions. Advertising can be of value when it is relevant to a user and not abusive of their attention or privacy. “Reverse metering” can credit users for the value (attention, data, content, and other) they provide to platforms and advertisers, offsetting any user fees. Negotiating the exchange on that basis might make it fairer, rewarding users commensurate with whatever level of attention and data they share — and make advertising more valuable to both consumers and advertisers.

Principle 4: Personal AIs delegated as faithfully loyal advocate agents

AI compounds the new sociotechnical problems that social media has produced but failed to solve. Providers of AI models exert similar, but far deeper and more subtle influence over what we see of the world and how we make sense of it. In addition to making parallel concerns of bias, manipulation, and misinformation less tractable to independent oversight, the seductive power of AI can threaten human attachment and lead from simple cognitive outsourcing toward cognitive surrender. As we move toward hybrids of human and AI intelligence, we face the same question: do our technology-based agents serve us, or the corporations that build and deploy them?

As we move from monolithic AIs to complexes of AI agents, the same pillars and need for open interoperability apply. For individual choice and agency: do we each decide what objectives are set and what sources are used? For social mediation: do we make full use of the wisdom of other individual humans “in the loop” and of the communities we choose? For reputation and trust: do we draw wisely on other humans and hybrids with transparency into who and what we are relying on?

AI developers have made much of the idea of growing agent capability, but not of AI relationship: OpenAI drew attention to capability as “agenticness,” the ability of an AI system to do things on its own. Richard Whitt and others emphasize a second dimension, which Whitt calls “agentiality:” an AI’s relationship to those it serves and whether it can be trusted to loyally serve them.

Given the complexity and dynamics of AI, our surest protection as individuals and communities is to be able to say “Have your AI call my AI.” That means selecting and delegating personal AI agents (PAIs) that we can trust to represent us as faithfully loyal advocates — sometimes even fiduciaries with legal duty to serve our interests, not those of their providers or any other party. Technology for ensuring trustworthiness is still at an early stage, but it is in active development.

Principle 5: The essential test question is “Who does it serve?”

Instead of reacting to the harms of misaligned technology, we should be building this brave new digital world to support fundamental and broadly supported goals. Many raise concerns about whose interests AI innovation is primarily serving. I suggest generalizing this into the single, simplest, most essential principle to keep top-of-mind as we design, develop, deploy, and use these new tools:

The essential test question at every decision point, for every component, is “Who does it serve?” This questions must be followed by two critical sub-questions: the normative “Who should it serve?” and the sociotechnical “Did we make sure it serves who it should serve?” All else is commentary.

We must make media tools intelligently social

We must look to the forest we want to live in, not just today’s nearby trees. Where do we want online tech to go? Our objective must be to make our media tools intelligently social — to fully support democratic freedoms and the flourishing of humans as individual-but-social creatures that cooperate in a full range of relationships, at multiple levels and in multiple domains.

To achieve that objective, we need a sociotechnical infrastructure that is broad-based, continuously evolving, open to diverse experimentation at a range of scales, extensible, resilient, and governed to serve each and all. The Pro-Human AI Coalition is well placed to catalyze that “big tent” effort, across both AI and social media.

We are re-engineering and digitizing — at unprecedented speed — social processes that humans have evolved over millennia. Humans and the AI systems that serve them will need 1) a sense of direction and 2) time to adapt to each other, shaping the tools as they shape us. In Part 2, I explore the roots and tools of each pillar, and how a loose, big-tent alliance can take the long view toward guiding and building that ecosystem as an ongoing, whole-of-society project.

Part 1 proposed five normative and architectural framing principles for a broad portfolio response to the risks and opportunities that AI and social media present to the public square, democracy, and human flourishing. Part 2 explores the historical foundation of this architecture and its continuing development, and it calls for a broad alliance to advance this momentous — and fraught — sociotechnical project for humanity's future.

The once and future process of our social mediation ecosystems

Human intelligence and sense-making depend heavily on our ability to communicate, cooperate, and learn from one another. Yet mainstream AI development often neglects this social substrate — as well as the related, and still poorly addressed, lessons of social media’s harm to the public square and democracy. Simplistic coding of social media, and now AI, lead to context collapse — losing the social context of epistemic dependence, norms, and values. Instead, digital augmentation needs compensatory attention to social context support. Meanwhile, current AI models collapse context even more.

Anticipation of AI agents has triggered promising research suggesting that, as agents interact with each other and with humans, they, too, will need the kinds of social institutions that human cooperation has always relied on. One study characterizes this as “the equilibrium-selection process through which multiple actors navigate their interdependence.” Another maps multiple levels of social mediation, across a range of scales and epochs, for “Existing Human Institutions” and for “AGI Institutions.” A thought-provoking analysis adapts to AI agents the way that human agents constitute democracy. One group of AI researchers offers a particularly evocative title: “Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt. Similar points had been raised from the social media perspective (with limited impact), such as Mediating Consent and Algorithmic Displacement of Social Trust.

This human process of collective sense-making, decisions, and action goes back to cavemen around campfires, refined through ancient Greece and the Enlightenment philosophers. The Founders baked this into American democracy, de Tocqueville observed how informal associations are vital workshops for democracy (1835-40), and Elinor Ostrom highlighted the collaborative wisdom of community institutions for governing the commons.

But now tech platforms are weakening the associations, communities, and civil society institutions that effective democracy depends on. This erosion began with mass media, as observed in Robert Putnam’s Bowling Alone. Social media, in turn, distorts the public square and offers no more than minimal support for community-centered interaction and governance. AI is deepening this failing, as “cognitive outsourcing” leads to “cognitive surrender” and “relationslipping,” and the “engagement economy” is driving deeper into our social being, creating a pseudo-social “attachment economy.”

The movement to decentralize social media is a powerful, but oversimplified, pendulum-swing reaction to this problem. The solution is a more nuanced recognition that society is built on dense multidimensional networks of communities and institutions that are semi-permeably connected — not just isolated, insular groups. Deeper openness in protocols and multi-agent interop enable vital architectural flexibility for rebuilding this equilibrium-selection process of democracy — human to human, AI to AI, and human to AI.

Social-media alternatives already demonstrate an increasing range of grassroots redevelopment through open protocols. Bluesky’s AT Protocol, for instance, now supports a wider “ATmosphere” of independent services, with Blacksky as an early example, and Eurosky and Gander emerging. Similar openness in AI could enable rich community-centered services and user-selected, community-curated attention agents.

Primacy of the agency pillar

Agency is the primal pillar of healthy discourse — necessary but not sufficient on its own. Unlike other social animals, we are defined by our human rights as individuals. We can delegate our agency to communities, tools, and agents (human or AI), but retain primacy.

Democracy’s central tension is between the interests of individuals and those of communities. Again, choice is essential to “cooperation: the equilibrium-selection process through which multiple actors navigate their interdependence” — even without any fixed hierarchy or other structure — and not an optimization problem with a “solution” to be imposed as some suppose.

Many fear excesses of agency, and some are unavoidable. But robust social mediation ecosystems and reputation systems nudge people toward responsibility and discernment. To live and thrive in any degree of harmony with the norms and values of society, one needs to maintain a good reputation. Done well, organic mediation enabled freedom of expression to flourish with little need for censorship (“moderation as removal”).

Open protocols are already becoming an alternative to reliance on monolithic social media platforms, enabling services that are more supportive of agency. The multiple platforms that make up the Fediverse and the newer, more flexible Bluesky ATmosphere ecosystem demonstrate growing demand for portability, open interoperability, and user control.

In AI, too, there is growing recognition that agency in choice and control is vital to make AI work in diverse domains and contexts, enabled by many dimensions of openness in models and sourcing. The challenges of effective cooperation of agents with humans and with each other has triggered a flurry of research on collaborative structures that has potential to reinvigorate both human democracy and how AI serves us. These challenges also add whole new dimensions to the principal-agent problem.

Tools for agency

The agency pillar has had the broadest influence in the tech and policy community, with much activity during the dot-com era, and now a resurgence in reaction to platform overreach and disruption. Recent social-media directions emphasize user-selectable feed algorithms and middleware, along with more control over personal data — but AI development is only just beginning to explore its equivalents, such as personal AI advocate agents. Legal mandates for relevant rights of delegation and fiduciary responsibilities, and for interoperation and portability, remain unsettled.

There is also growing recognition that monolithic AI models are unreliable (and opaque) about the sources they draw on, which makes user-controlled sourcing essential to quality results. Many businesses are building domain-specific AI models, but a more general solution may be in models that are designed for Attribution-Based Control (ABC). This is very analogous to the source selectivity in social media feed algorithms that social media middleware seeks to enable.

Social mediation: tools for communities as first-class institutional principals

Communities need tools that serve them as institutions, with agency of their own — not merely as sets of users. Social mediation may be the least recognized and understood pillar of healthy “AI/SM” (AI and social media as one), even as much current prosocial work is rediscovering it. Goldberg and her 26 co-authors make a compelling case for developing together the related families of technology that support social mediation: deliberation, bridging and community-driven moderation, with proof of humanity as a foundation for all three.

Goldberg et al. present a compelling call to better serve communities, including by public-centered design, multilevel governance, and bespoke tooling, with support from civil society, academia, and the public sector complementing industry.

I suggest taking empowerment further: to give individuals and communities structural power to choose, substitute, govern, and federate systems and agents. This is where agency and social mediation intersect. Because this technology is mission-critical, specialized, and context-specific, communities should draw on the long-established example of “enterprise IT” to be able to select and customize their own tools and loyal advocate agents from an open market, rather than depend on a general-purpose platform provider.

Communities need tech tools that their leaders, staff, and agents (human, AI, and hybrid), can use and manage to work with their members. Examples include member relationship, content management, community governance, federation, and inter-community negotiation systems. They may come from a mix of for-profit, non-profit, and public suppliers. This requires a supporting ecosystem of “Independent Software Vendors (ISVs)”, system integration services (across families and pillars and operational support), modularity, open tech stack support for a full spectrum of institutional formality, size, and scope, and inter-community services for managing graduated degrees of permeability (and safety).

Digitizing systems support for the reputation/trust pillar

Reputation and trust online need far more systematic support than the word of mouth they have traditionally relied on. Goldberg et al. address proof of humanity as a “foundational need” that applies across the other three technology families as increasingly critical infrastructure for online interactions and recognizes a spectrum of hybrids. The authors also discuss reputation systems as relevant to community-driven moderation.

I generalize and distinguish reputation as a third broad trust pillar, including proof of humanity and much more, that is dynamically derived from, and feeds back into, the processes of the other two pillars.

Traditionally, reputation and trust were intuitive and informal, drawing primarily on word of mouth. In some domains they grew into more structured and institutional “cred”, such as citations, reviews, ratings, and professional standards. Some social media services have had success with simple systems for social reputation, such as “karma" in Slashdot, then Reddit, but there is unmet need for much more — to shift from an attention economy prioritizing empty engagement to one prioritizing trust.

Because trust varies by community, domain, and task — and changes over time — reputation systems should be contextual and constantly updating. They should be categorized to apply to people, communities, content, and AI agents — including tracking of pseudonymous participants and human-AI hybrids. To address varying perspectives on authority, they should flexibly combine bottom-up judgments with context-relevant institutional credentials.

Reputation is a dynamically updating attribute used by the other two pillars. Like all components, it depends on supporting technologies for integrity, security, and privacy.

Tools for reputation/trust

Trust can be built from the bottom up, without reliance on any one central authority, much the way Google's PageRank ranks web pages: by trusting those trusted by those we trust. Goldberg et al. note use of PageRank as a trust signal, and use of reputation systems in community governance. I generalize these into a pillar that can leverage scalable, user-selectable, and portable systems, for bottom-up and transitive trust — rather than rely only on ratings, “karma,” or centralized authority.

Personalized PageRank-inspired methods can compute the transitivity of trust by degrees of separation (trust those trusted by those I trust, n-levels deep), as I have elaborated. This is distilled systematically from a global population, but applied with respect to specific sources and domains. It can be protected against gaming (albeit as a continuing cat and mouse game), and reduces the need for some external “authority” whose legitimacy and contextual knowledge are open to challenge.

Open interoperability: a flexible enablement foundation

Open interoperability, both vertical and horizontal, is the foundation that enables diverse elements of all three pillars and other elements to work in concert — and be changed as needed. Protocols and APIs already support complex business enterprise systems. They remain underdeveloped for individuals and communities, but are beginning to emerge in social media and multi-agent AI.

Interoperability enables competition, portability, and “exit” without losing data or connections — allowing old and new components to be mixed, matched, substituted, and extended. That flexibility has costs, including additional security demands, but monolithic, proprietary systems are too rigid to accommodate the diversity and ongoing evolution of human society.

Dominant platforms lack strong incentives to support the pillars, or the interoperability that would enable open support for them. Interop can weaken their control, since public-good support for the pillars may not directly advance advertising revenue. That incentive problem is an important obstacle for society to address — here again, a portfolio of innovative models should be pursued.

Other technical, resourcing, and civic foundations

Beneath the pillars and interoperation sit three other kinds of foundations: technical safeguards that every component needs, resourcing and incentives, and the policy and civic support that markets alone can't provide. Widely accepted technical foundational needs include alignment, integrity, security, privacy, and transparency (legibility). These are basic requirements that underlie all three pillars.

Many current problems are recognized as related to resourcing and incentives, but that derives primarily from an agency problem. Again, a portfolio should be pursued: negotiating attention and data sharing for advertising, potentially based on reverse metering; as well as alternative resource sources, including both fee for service and public support. To be fair, pricing should be value- and outcomes-based.

At the sociotechnical and civic level, market demand may advance build-out of the pillars, open protocols, and AI agents, but policy, regulation, and standards could greatly facilitate securing agency, delegation, interoperability, portability competition, and public-interest infrastructure. US policy has badly neglected this, but the 2026 midterms are catalyzing a new energy. Broad civic education is also needed so the public can understand technological harms and demand remedies aimed at root causes rather than symptoms — and demand that of their government representatives.

A loose, big-tent, multidisciplinary, whole-of-society project portfolio

AI and social media are merging but being addressed separately, and neither is being built to support the communities essential to our humanity. Closing these disconnects will take a broad alliance working from shared principles, and the Pro-Human AI Coalition seems well placed to start that.

How we govern AI/SM will determine 1) whether it augments or impairs collective intelligence, cooperation, and the workings of democracy which already face growing challenges; and 2) whether humans can keep governing and controlling increasingly powerful and potentially dangerous AI agents. Because no one pillar can stand alone without serious weaknesses — and given the very different challenges in social media and other uses of AI — shared framing principles can provide common understanding, objectives, criteria, and motivation for advancing a richly diverse interoperable and adaptable portfolio within, across, and below the pillars.

Only through cooperation on shared goals can we develop a coherent portfolio approach to marshal the tools and to guide the long-term project of sociotechnical evolution. We need to re-engineer democracy for the digital era and to govern AI — to overcome impending threats and enable humanity to flourish. There are organizations supporting work in this direction, but cooperation remains less integrated than needed across the pillars and across AI and SM. A big-tent alliance could make that broader alignment visible and actionable, and could refine and build on framing principles like those I propose here. For example, it could explore how my Three Pillars relate to the framework in Figure 3 of Weyl and colleagues’ Prosocial Media paper, and how these relate to stated objectives of the coalition, as I discuss in the Supplement.

Project Liberty’s longstanding work on the social web gives the new coalition a promising opportunity to integrate the social media dimension into its agenda. A practical next step would be for this new coalition, or others, to workshop shared framing principles and improve information-sharing across a loosely federated portfolio. That could support diverse experimentation and ongoing evolution of how the pillars and other elements work together interoperably, without imposing a single program.

Conclusion and call to collective action

We are adding AI and social media into a society that evolved over millennia to rely on communities and institutions for cooperation, discourse, sense-making, and democratic governance. Yet development remains piecemeal and concentrated in corporate systems whose incentives do not reliably prioritize long-term public value — and more pro-human re-engineering remains mostly reactive, with little overarching vision. Better alternatives are emerging, but remain obscure, fragmented, and constrained by platform structures and extreme concentrations of power. Current directions in AI seem to be continuing the neglect of human communities and institutions. Humanity cannot survive and flourish in a “post-social” world of people talking to computers — we need a world of people talking to people, augmented by computers.

The proposed principles address five questions: 1) how thought remains a social process; 2) how agency, social mediation, and reputation reinforce one another when enabled by open interoperability; 3) how users regain control of their attention and data, 4) how personal advocate AIs can help individuals and communities ensure AIs that serve other parties respect their agency and rights; and 5) above all, how every component must be judged by who it serves. These principles should be simple and clear enough to be understood and demanded by all.

The goal is to make our media tools intelligently social—supporting democratic freedom and human flourishing through an open, evolving ecosystem. Building it requires a broad alliance that can coordinate this sociotechnical evolution without suppressing diversity and subsidiarity. The Pro-Human AI Coalition could pursue this by convening its members to workshop these shared principles — and drawing all of us into shaping this momentous transformation of society.

Acknowledgements: Thanks to my previous co-authors and editors on these themes and to the many other colleagues for key insights and valuable feedback that I have built on. And thanks to Audrey Hingle for insightful editorial support. (AI was used for supplementary research and analysis, and for suggesting condensations of text that I first drafted and then carefully checked word by word.) This piece expresses my personal views, not necessarily those of any organization I am affiliated with.