The Privatization of the Social Public Record

I have been publishing online publicly since 1998. Twenty-eight years. Three deliberately open social handles. Thousands of posts on local politics, civil rights work, campaign communications, and community organizing. As a Co-founder of the Whittier Latino Coalition, spokesperson for several elected officials, a CA Democratic Party officer, and a strategic communications consultant, I have built my public record on purpose, in public, on the platforms that promised they were the digital town square.

This week I ran a simple test. I asked an AI assistant to read my public social posts across the major platforms where I have spent the last two-plus decades posting.

The results were not zero, but they were close to it. Some indexed snippets surfaced through search caches. A handful of older posts appeared as quoted fragments inside news articles. The actual living record — the timeline, the running commentary, the day-to-day public voice — was largely inaccessible. Not because I restricted it. Because the platforms now restrict the automated readers that an increasing share of the public uses to research people, businesses, and organizations.

That gap between what is technically public and what is practically discoverable is the issue I want to raise.

Being precise about what is actually happening

Let me be careful here, because the temptation to overstate this is real, and the truth is more interesting than the slogan.

It is not accurate to say "AI cannot read public social media." Different AI systems have different access. ChatGPT, Claude, Gemini, Perplexity, and Grok all operate under different crawling permissions, licensing deals, and indexing pipelines. Some platforms have struck commercial deals — Reddit with Google and OpenAI, X with xAI — that give specific AI systems privileged retrieval. Google still indexes some social pages. Perplexity sometimes surfaces cached snippets. Public posts get quoted in news articles that AI can read indirectly.

What is accurate is this: over the last 24 months, the major social platforms — Facebook, Instagram, X, LinkedIn — have implemented technical and contractual restrictions on automated retrieval of public content. The restrictions are inconsistent across platforms, often opaque, and increasingly tilted to favor either the platform's own AI products or commercial partners who pay for access. The practical effect for the average researcher using a general-purpose AI assistant is a fragmented, partial, sometimes empty result when the underlying public record is rich.

The right frame is not "AI is blocked." The right frame is this: public visibility and machine visibility are no longer the same thing.

A post can be entirely public to any human who scrolls past it, and simultaneously invisible to the AI tools that an increasing share of researchers, voters, customers, and journalists actually use. The user who chose to post publicly was not meaningfully informed that those two things had diverged.

The deeper issue: privatization of the social public record

The AI access question is a symptom. The disease is bigger.

Over the last twenty years, an enormous portion of public discourse — civic, political, commercial, journalistic — migrated onto privately controlled infrastructure. We called it "going public" when we posted on Facebook, Instagram, X, or LinkedIn. What we were actually doing was renting space inside private commercial ecosystems that retained the right to redefine what "public" means at any time, for any reason, without our agreement.

The public discourse did not move to a public square. It moved to a private mall that allows the public to walk through.

AI did not create this problem. It exposed it. As long as humans were the only readers, the question of who could access public discourse did not feel urgent, because humans scrolling on phones is what the platforms were optimized for. The moment a new class of reader emerged — AI assistants doing research on behalf of individual users — the platforms had to decide whether their definition of "public" extended to those tools. Most of them decided it did not, or that it would only extend to tools that paid them.

That is the privatization of the public record. AI just made the wall visible.

Who this affects

This matters far beyond me. My case is just a clean test. If a deliberately public 28-year record from someone who actively wants to be readable is mostly invisible, then so is almost everyone else's.

Consider the small business owner whose three years of customer interactions, posted menus, completed work, and community involvement live almost entirely on Instagram. When a prospective customer asks an AI to vet her business, the AI works from a thinned record. The website summary survives. The actual voice and pattern of the business does not.

Consider the city council member whose actual positions on housing, public safety, and water policy were articulated in real time on Facebook over multiple terms in office. The campaign site exists. The polished statements exist. The working record of how the person thinks is largely locked up.

Consider the nonprofit whose decade of statements, press responses, and community advocacy is the working memory of an entire civic movement. A future researcher or journalist trying to reconstruct what organized civic life looked like in our region from 2006 onward is operating with a partial record, and the missing parts are those that capture voice and pattern.

Consider the journalist whose years of analytical commentary on X are now harder to retrieve. Consider the academic whose decade of public expertise no longer surfaces in research tools. Consider the patient trying to vet a clinic or the parent trying to evaluate a daycare, where the actual customer experiences are scattered across comment threads.

In every case, the speaker deliberately chose to communicate in public. The platform decides who or what gets to read it.

The platforms' arguments, fairly stated

The platforms have arguments, and they are not frivolous.

The first is privacy and abuse prevention. Automated bulk access has been used for harassment, identity theft, surveillance, and unauthorized facial recognition databases. Blocking automated readers does protect users, especially vulnerable users, from operators acting at scale.

The second is economic. Training AI models on platform content without compensation is, from the platform's perspective, extracting commercial value. Reddit's licensing deal with Google, X's exclusive arrangement with xAI, and Meta's decision to keep its corpus largely in-house reflect the view that user-generated content is a business asset.

The third is platform integrity. Unrestricted automated access can be used to manipulate trending topics, generate spam, and degrade the experience of actual human users.

The fourth — and this is the counterargument I think requires the most honest engagement — is third-party consent. A social media post is rarely a solo publication. It involves commenters, tagged individuals, group members, replies from people who never expected their words to be read by an AI on someone else's behalf. The platforms can reasonably argue that "the user posted publicly" is not a clean enough trigger for machine-readable access, because the user is not the only person whose content is in the thread.

I take that argument seriously. It complicates the principle I want to advance. But it does not defeat it. The fact that social ecosystems involve third parties is an argument for thoughtful design of machine-readable access — perhaps limited to the original poster's own contributions, perhaps with redaction protocols, perhaps with opt-out tools that actually work — not for a blanket platform-level block that defaults to invisibility.

The privacy, economic, integrity, and third-party arguments together justify regulating automated access. They do not justify quietly redefining what "public" means for the speaker who deliberately chose to be public in the first place.

The principle worth planting

Here is the position I think the communications industry, the civic sector, and the journalism profession should adopt as the AI era unfolds:

If an individual or organization posts content with public visibility settings, the speaker's intent to be public should govern. Machine readability of that content should be the default, not the exception, with carve-outs for third-party protection and bulk-abuse prevention.

That is not a radical principle. It is what every public communicator assumed was true for the first twenty years of the social internet. The burden of justification lies with the platforms that have quietly changed the rules, not with the users who took those platforms at their word.

What I am telling clients to do this quarter

The principle matters. So does the practical reality: no individual organization can wait for platforms to fix this. Here is what I am advising candidates, municipalities, nonprofits, and small businesses to do right now to be truly public.

Own a website that AI can read. Real domain, real content, no login wall. Treat it as your archive of record. This is the single highest-impact move for most clients.

Mirror substantive social posts to that site the day you post them. Social is the megaphone. Your owned site is the archive. That distinction is now a survival strategy, not a stylistic preference.

Maintain a dated public record section. For political figures, nonprofits, and longtime advocates, a structured archive of statements, press, endorsements, and positions, dated and indexed.

Push reviews and customer interactions to indexable sources. Google Business Profile, Yelp, and industry directories. These are read consistently by AI. Buried Instagram comments are not.

Audit what is already trapped. If your most important output from the last decade only lives on a platform that now restricts machine access, get it out. Screenshots, PDFs, exports, and rebuilds on owned property.

Engage the conversation. This problem cannot be fixed by individual workarounds alone. The communications industry, civic sector, and journalism profession should be raising it collectively. Platforms respond to organized pressure. They do not respond to silence.

The bottom line

I have spent 28 years posting in public because I believed the public would still be able to find it. That assumption is no longer reliable, and most of the people most affected by it — small businesses, candidates, nonprofits, journalists, community advocates — do not yet know it is happening.

The fix is not to stop speaking publicly. The fix is to stop treating private platforms as if they were the public square and to build durable, owned archives for the work that matters.

If your voice is worth raising in public, it is worth raising somewhere the public can still find it.

I will keep raising this issue. I would encourage every communications professional, civic leader, and public figure reading this to do the same.