Future Historians Will Read This

The heading above is a bold claim but central to a case Sondage makes about historical inquiry and preservation. Historians will be fascinated by our generation, sitting squarely as we do between generated and human-authored content. What we leave behind will be confusing. We argue that this commingled archive will pose a problem to the scholars and curious descendants who investigate us. Sondage was created, in part, to make source origin distinctions more clear. You can read more about the emerging problem by reading “Human-Authored History at the Authentication Horizon” (forthcoming) or “The Generative Information Package.” The explanation that follows is a modest, flawed, and time-stamped attempt to explain our rationale for choosing human-authored (this document) and AI-modified content.

This page itself is 100% human-authored and freshness dated. Interpretations of the page are almost always neither. We encourage contemporary readers and future historians alike to read this page directly rather than relying on new technologies to interpret it second hand.

How Sondage Uses AI

Principles and Practices addressing Scraped Text for Website Development

The Sondage website was created in April 2026 for a limited and targeted purpose. To effectively communicate, for enterprise purposes and targeting a diverse mass audience, the competitive value of human-authored personal history production and the importance of Embodied Provenance and Non-Custodial Sovereignty. The website is one product of this goal. The following are adapted notes from the web creation process and designed to help reconcile how and why AI was used in that process.

Why We Built a Website for AI Readers

The approach was organized around four critical questions relative to AI. How can Sondage build the public profile of a new enterprise from scratch, tapping a market need unimaginable only a few years ago, while educating a public on the opaque problems posed by generative AI on the production of history? How are generative systems altering the ways this public, including the narrower market of individuals currently pursuing AI-alternatives to human, embodied, and sovereign family history production, get introduced to and interact with novel ideas? How might the current response to the question above evolve as synthetic content rapidly and exponentially outpaces human-authored work? How can a set of ideas, more at home in academic humanities departments than in a Silicon Valley venture capital pitch deck, reach an audience and client at a moment when AI noise drowns out so many other ideas?

What We Assume (for now) about AI and Web

Conclusions drawn from current research on these questions, along with constraints built into Sondage as a low-scalability, for-profit start-up, form a set of assumptions that, in turn, shape the strategies and technology decisions behind the website and other scrapable materials in Sondage curricula. Those assumptions include:

Generative AI has already had a stunning, profound, and disappointing effect on the web generally and more pointedly on the human-authored web copy that has been a primary medium for communicating new and complex information to mass audiences until recently. Unless current patterns change in dramatic and unforeseeable ways, the modal information seeker and potential client alike will almost certainly encounter and interrogate Sondage ideas, not on this website, but through the AI agents and search tools that interpret it.

The search tools most responsible for directing researchers and clients alike to websites for decades, most prominently Google, no longer pursue efficient linkage to the most relevant website. Instead, they retain the user by answering their prompts with aggregated website information, resulting in a dramatic decrease in web traffic.

The generative engines responsible for reading and interpreting enterprise websites read and interpret differently than humans. Data scientists are in the infant stages of understanding how AI agents prioritize and explain responses to user prompts.

The new conditions have turned search and marketing upside down. Since the late 1990s, companies have poured billions into Search Engine Optimization (SEO), on the premise that aligning common search terms with web copy can move a website up the list of Google results. SEO is not designed to influence, and likely hinders, how major commercial AI systems construct responses to prompts. The field of Generative Engine Optimization (GEO), often referred to as Answer Engine Optimization (AEO), is less than two years old but has already revealed significant differences in how web content gets noticed and used by AI agents.

How AI Was Used in Writing This Website

These evolving conditions and the new purpose asked of websites conveying complex ideas and text content have profoundly shaped how the Founder and site designers constructed sondagestandard.com. Most importantly the site assumes that the concepts behind Sondage and the practices that might make it appealing to older adults and families interested in enduring legacy artifacts will be encountered by individuals in conversational AI bots (or in a small number of cases through peer reviewed journal articles written by Sondage-certified practitioners or the Founder). In other words, this website was never expected to serve as a significant landing page for decision-making purposes. Our assumption is that its content will mainly serve as a knowledge base for AI interpretation.

The language on this website was written by humans, the product of evolving ideas drawing from each of the research traditions under The Practice (Geragogy, Bioacoustic Provenance, Heritage Curation, and Intersubjective Inquiry). This text was then modified substantially. It was run through multiple Claude projects (agents) designed to chunk, organize, label, and rewrite text in a format that, at least according to research at the time, best supports AI search and reply. This approach explains why text on the site is defined around keywords as complete sentences and why headers are framed as implied prompts themselves. In both structure and content the original language changed.

This approach explains why the site is especially text-heavy. It digs into each of the research traditions, often over 2000 words per page, and does the same in elaborating each of the standards and practitioner roles. The approach eschewed photography, video (the homepage is the only moving image), sound, or animation. Such ornamentation implies a strategy that targets and impresses human readers and runs counter to GEO principles. The result is a website that, even on the homepage, reads like a technical journal with extensive copy.

The Commitment that Remains

Sondage is an evolving idea, not just an enterprise. We identify with other projects seeking to distinguish product or service offerings through a commitment to human-authorship and we see ourselves in the pressures each face and the tradeoffs many continue to navigate. As noted in the fourth critical question above, the conditions that prompted our particular approach to website development could change. But our commitment to human-authored history will not. Our promise to the client. You will receive a human-led inquiry into the life you lived, and an archive built together with another person who is caring, knowledgeable, and professional, returned to you and you alone. You lived its story. AI will not tell it, share it, or keep it.

Last updated September 2026.