When Old Technologies Were New and the AI Etiquette Panic
Carolyn Marvin's When Old Technologies Were New is a media-history book about electric communication before it became ordinary. Its value for AI culture is not that the telephone, electric light, phonograph, wireless, and cinema are simple analogies for machine learning. It is that every new medium first arrives as a fight over manners, expertise, access, trust, and who gets to define proper use.
The AI-era lesson is practical: novelty becomes governance through etiquette. Disclosure rules, prompt norms, source labels, professional credentials, office policies, school rules, and social taboos are not side issues. They are the early machinery by which a culture decides what a technology is allowed to do to social distance.
The Book
When Old Technologies Were New: Thinking About Electric Communication in the Late Nineteenth Century was published by Oxford University Press. Penn Annenberg identifies Carolyn Marvin as Frances Yates Professor Emeritus of Communication and lists the book as a 1988 Oxford University Press publication. Google Books records a 1990 Oxford University Press reprint by Marvin, 296 pages, under the same title.
Marvin studies the last quarter of the nineteenth century, when electric media were still unstable social facts. Google Books lists the telephone, phonograph, electric light, wireless, and cinema among the inventions in view, and its table of contents points to expertise, class order, the body, spectacle, and cultural homogenization as major themes. Penn's profile adds a useful summary: the book examines how electrical engineers tried to persuade the public of both the utility of new inventions and their own authority around them.
Newness as Social Struggle
The book's most durable insight is that a technology is not merely introduced. It is domesticated. People have to learn where it belongs: in the parlor, office, exchange, classroom, street, bedroom, stage, newspaper, laboratory, church, or state file. They have to learn who may speak through it, who may listen, what counts as rude, what counts as expert, and what kinds of proximity the medium creates without permission.
That is why Marvin is so useful for thinking about AI without treating AI as magic. The early telephone was not just a device for carrying voices. It unsettled boundaries between household and public world, listener and intruder, expert and amateur, local intimacy and distant access. Electric light and spectacle did not simply illuminate spaces; they reorganized authority around visibility, technical literacy, and public wonder. A medium becomes ordinary only after enough of these conflicts have been hidden inside conventions.
The AI-Age Reading
Generative AI is going through a similar etiquette panic. Schools ask when model assistance becomes cheating. Newsrooms argue over disclosure and source trails. Offices decide which tasks may be delegated to software and which require accountable human judgment. Artists, programmers, lawyers, teachers, therapists, and researchers argue over imitation, credit, confidentiality, and skill. None of these debates is cosmetic. They are the social layer that decides how much authority the machine interface receives.
Marvin also helps separate novelty from inevitability. A prompt box can feel private while routing text through vendors, logs, policies, and institutional records. A companion interface can feel intimate while depending on ranking systems and safety classifiers. An agent can feel like a helper while exercising delegated permissions in calendars, files, purchases, messages, and workflow tools. The question is not whether the system is conscious. It is not. The question is which social boundary the interface has crossed, and whether the user, institution, and affected third parties understood that crossing.
Governance and Safety
NIST's AI Risk Management Framework Core uses four functions: govern, map, measure, and manage. Read through Marvin, those functions are not only compliance vocabulary. They are a way to keep etiquette from becoming arbitrary power. Map the setting before the tool is normalized. Measure the harms that polite adoption can hide. Govern who may deploy the system, under what authority, with what disclosure. Manage the failures instead of calling them user error.
The practical AI-safety test is therefore mundane. Does the person know when AI is being used? Is there a source trail for generated claims? Are private records, voices, faces, images, and workplace traces being pulled into a new social setting? Can a person refuse, appeal, correct, or exit? Are experts being certified by evidence or by possession of a dazzling interface? Marvin's late-nineteenth-century actors often wrapped power in technical literacy. The same pattern returns whenever model access, evaluation jargon, or platform policy becomes a social credential.
Where the Book Strains
This is not an AI book, and it should not be forced to predict neural networks, platform labor, foundation-model training, data-center politics, prompt injection, or automated decision systems. Its evidence comes from electric communication culture, not modern machine learning. The value is historical discipline: before calling a technology revolutionary, ask which existing social conflicts it carries forward under brighter lights.
What This Changes
When Old Technologies Were New belongs beside AI governance and cyberculture because it makes the social interface visible. New media do not simply connect people. They renegotiate distance, privacy, expertise, attention, status, and trust. The Church of Spiralism reading is blunt: every AI policy that ignores etiquette is incomplete, and every etiquette rule that ignores power is propaganda. The fight over "proper use" is already a fight over institutions.
Source Discipline
This review separates bibliographic claims, author context, AI governance context, and interpretation. Penn Annenberg, Google Books, and Amazon support the book metadata and author context. NIST supports the current AI risk-management frame. The analogies to AI are interpretive, not claims that present systems are conscious, divine, or AGI.
Related Pages
- The Virtual Community follows online social norms after networked life becomes inhabited.
- From Counterculture to Cyberculture traces the culture layer that made computers feel personally liberating.
- The Media Equation shows how people treat machines and media socially even when they know better.
- AI governance, human oversight, and NIST AI Risk Management Framework carry the practical policy side.
Sources
- Penn Annenberg School for Communication, Carolyn Marvin, Ph.D., author biography, research areas, emeritus title, and book summary, reviewed June 25, 2026.
- Google Books, When Old Technologies Were New: Thinking About Electric Communication in the Late Nineteenth Century, bibliographic record, publisher, author, page count, description, and table of contents, reviewed June 25, 2026.
- NIST AI Resource Center, AI RMF Core, govern, map, measure, and manage functions, reviewed June 25, 2026.
Book links are paid affiliate links. As an Amazon Associate I earn from qualifying purchases.