Proof of care in the age of AI

In a world where a polished 4,000-word essay can be generated in ten seconds flat, the long-standing equation between effort and meaning has collapsed — and proof of care has become one of the most urgent unsolved problems of our time. The question is no longer whether a piece of content is well-written; it is whether any human actually cared enough to produce it. In his widely-shared 2026 essay Proof of Care in the Age of A.I., Jacob Filipp argues that AI's greatest disruption may not be economic but semiotic: it has degraded a crucial signal that humans have used for centuries to establish trust, connection, and persuasion. It is worth flagging up front that Filipp's piece is a speculative essay — it mixes genuine present-day advice (handwrite your posts; keep tattoos short) with deliberately exaggerated near-future worldbuilding, complete with invented settings such as "Designated European Warzone II" and "the Euphrates Crucible." Read as a thought experiment rather than a news report, it crystallizes what confronting proof of care in the age of AI actually demands: it may be the defining authenticity challenge of this decade.

This is not a niche concern for content marketers or academics. It touches every layer of communication — personal relationships, political speech, journalism, commerce, and culture. Understanding the mechanisms people are already using, and the more extreme ones Filipp imagines, is a useful lens for anyone thinking seriously about what a human-centered digital future actually looks like.

The Signal That AI Broke

Before large language models became a commodity, effort itself functioned as an implicit certificate of authenticity. When someone sat down and wrote a long, considered piece — on Pomeranians (Filipp's own example), on grief, on economic policy — the sheer labor involved acted as a credibility signal. It was costly to produce, so production implied genuine investment. Economists call this a costly signal: its value derives precisely from how hard it is to fake.

Generative AI has not merely reduced the cost of writing. It has abolished the cost asymmetry that made effort meaningful. When a diligent human author and a careless one with a chatbot can produce indistinguishable outputs, the effort signal carries no information. Readers, recipients, and audiences face a verification problem they have never encountered at this scale before.

Why it matters: The collapse of effort as a signal does not just affect content quality — it degrades the entire social infrastructure of trust. Persuasion, intimacy, and civic discourse all depend on the ability to sense that someone genuinely cares about what they are communicating. When that sense is gone, the default shifts toward suspicion.

Filipp identifies a sharp irony here: AI was supposed to make it easier for people to put their thoughts into writing. And it has. But that ease is precisely the problem. A tool that removes friction from authentic expression simultaneously removes friction from inauthentic expression, and humans have no reliable way to tell the difference from the output alone.

Emerging Methods for Proving You Actually Care

Rather than waiting for a technological solution — a universal AI-detection tool, a cryptographic provenance system — Filipp imagines people and communities improvising analog and social proof-of-care mechanisms. Some are things he presents as already practiced; others are deliberately escalated hypotheticals he uses to probe where the logic leads. They range from the intimate and obvious to the startling and extreme. Each comes with its own logic, its own corruption vector, and its own shelf life.

Handwriting: Costly Signals at Human Scale

The most intuitive response to digital abundance is to return to handwriting. The physical act of writing by hand cannot be meaningfully automated at the personal level — no shortcuts, no autocomplete, no copy-paste. When someone hand-writes a social media quip, photographs it, and shares the image, the medium itself carries the message: I cared enough to do this the slow way. Filipp argues this "will hit home better than if you'd written it digitally."

Filipp describes a more friction-intensive variant: manually mirroring or reflecting each letter, producing text that is harder to read at a glance but conspicuously deliberate in a culture of compulsive skimming. He even argues that tracing AI-generated text by hand carries some proof-of-care weight — the physical commitment of rewriting, letter by letter, still signals a level of engagement that copying and pasting simply does not.

At a larger scale, Filipp suggests hand-writing and distributing physical flyers retains significant power precisely because it does not scale easily. The key discipline is extreme brevity: anyone writing a thousand copies by hand quickly learns to cut every unnecessary word, or risk genuine wrist injury. That constraint produces a different kind of communication — dense, memorable, and visibly human.

The Pen Plotter Problem: When Signals Get Spoofed

No effective proof-of-care signal survives forever without being counterfeited, and in Filipp's telling handwritten flyers are no exception. His scenario imagines wealthy and organized interests co-opting pen plotters — machines that mechanically reproduce handwriting with uncanny accuracy, including deliberate imperfections and one-of-a-kind fonts modeled on real individuals' handwriting styles. In the essay, cartels of businesspeople hand plotter-produced flyers to gig workers who distribute them in town centres, manufacturing the appearance of organic, grassroots communication at industrial scale.

This is presented as a near-perfect attack on the handwriting signal: the output looks hand-made; the process is automated and coordinated. Within Filipp's imagined timeline, "over the past three years the public had smartened up to the use of plotters," so handwritten flyers come to be viewed with more suspicion than they once were — which, in the story, drives the search for new, harder-to-fake indicators of genuine human effort. The arms race between authentic signals and their sophisticated imitations is the recurring theme across every method he discusses, and it sits at the heart of the broader proof-of-care problem in the AI age.

Tattoos: Permanent Commitment as Broadcast

A tattoo is, among other things, a permanently expensive signal. The pain of acquisition and the irreversibility of the outcome together constitute a credibility mechanism that is hard to fake and impossible to retract. A message inscribed on someone's body is on passive broadcast to everyone they encounter for the rest of their life — no ongoing effort required after that initial commitment. Filipp notes it works "for as long as real tattoos can be distinguished from the fake thing."

The constraints are severe by design. The message must be short (skin is finite) and evergreen (positions change; tattoos do not). Filipp cites the cautionary examples of "Meat is murder" tattooed down the middle of a young woman's torso and "Grind Hard" — both perfectly legible proof-of-care signals at the time of inscription, both plausibly awkward decades later. The very permanence that makes a tattoo credible also makes it a high-stakes commitment to the content of a belief, not merely to the act of holding one.

The Imagined Revival of Oral Culture: Town Criers and Travelling Storytellers

One of the more sociologically striking moves in Filipp's essay is a hypothetical revival of in-person public communication. In his scenario, more people take to actively shouting messages in urban spaces — a form of town crying that counteracts the perceived misinformation density of online channels by anchoring speech to a physical, identifiable human body. Presence itself becomes the credential. It is important to be clear that this is Filipp's speculative framing, not a measured, documented trend backed by data — there is no evidence of any actual real-world revival of town criers, and this article makes no such claim.

More elaborate still is his imagined emergence of travelling storytellers: individuals who move between cities, sit with local communities, and share narratives adapted to the specific identity and energy of each audience. Office workers, warehouse workers, and nurses respond to different framings of the same underlying story. The storyteller's craft is precisely this adaptability — a skill that is labor-intensive to develop and hard to fully automate, because it requires real-time reading of a room.

Filipp reaches for two historical touchstones to explain the appeal. He writes: "I hear that this is how the Iliad and Odyssey were first transmitted" — by travelling performers who adapted the epics to local audiences — "except now, it is more like Julius Caesar's autobiographical books: each story is not just entertaining but aims to telegraph a message about a person or initiative." In other words, eyewitness narrative attributed to a specific named body in a specific named place can carry a persuasive authority that anonymous text cannot. The storyteller as propagandist is an ancient model; Filipp's twist is to imagine its revival in an AI-saturated information environment as a logical, if unsettling, adaptation of the same insight.

Guilds and Physical Initiation: Anti-Corruption Infrastructure (Filipp's Thought Experiment)

If storytellers can be hired by powerful interests to spread coordinated narratives, their proof-of-care value collapses — just as, in Filipp's story, it did with pen-plotted flyers. His imagined response is the formation of storyteller guilds that use deliberately painful physical initiation rites to establish and verify authenticity. This is where the essay tips fully into speculative worldbuilding, so it should be read as a hypothetical rather than a report of anything actually observed.

The logic parallels that of tattoos, but is collective and procedural rather than individual. In the scenario, male initiates undergo skin cuts made with razors, with burnt wood ash rubbed into the wounds, healing into raised, alligator-scale scarring; female initiates have ornate shapes carved into the face using a bone chisel. The resulting marks are meant to be visible and verifiable: a listener could assess a storyteller's guild membership at a glance, without trusting any digital credential or institutional certificate. The guild scar functions, in effect, as a biological analog to a blockchain — a tamper-evident ledger written on the body.

This imagined model is a radical departure from how digital systems approach trust, where verification is cryptographic and invisible. The guild model makes verification sensory and social, which is precisely what — in Filipp's argument — would make it resistant to the kinds of spoofing that plague digital credentials.

The Escalation Problem: Where Does This End?

Filipp's most sobering thread concerns escalation dynamics. Each proof-of-care signal, once sufficiently well-known, faces pressure to be topped by something more extreme. Handwriting gets spoofed by plotters; scarification gets adopted by organized actors; each ratchet upward demands a more costly, more irreversible demonstration of commitment.

The logical endpoint of this escalation, in his telling, is body subtraction — literally removing parts of oneself to underscore a declaration. Here the essay reads openly like dispatches from a near-future dystopia, set in invented locales: businessmen severing an ear to seal a high-stakes deal; government ministers cutting off a finger when formally declaring war in fictional conflict zones such as "Designated European Warzone II." These are Filipp's imagined illustrations, not documented events, and no such practices exist in reality. The egalitarian appeal he sketches is real — unlike a tattoo studio or guild initiation, amputation requires no external institution — but the escalation logic is deeply troubling.

Filipp explicitly traces this trajectory toward self-immolation, invoking Tibetan protesters of previous decades as a real historical example of how escalation dynamics in proof-of-sincerity can reach an ultimate terminus. His point is not that everyone will start cutting off ears, but that without stable, institutionally-maintained proof-of-care mechanisms, informal escalation pressure could push toward increasingly extreme demonstrations, with no natural stopping point. This is the darkest implication Filipp draws about proof of care in the age of AI: a technological disruption that began with text generation could, taken to its extreme, intensify the most ancient and visceral human rituals of commitment. (Fittingly, Filipp undercuts his own dystopia with a wry closing note — "The kids are allright" — a reminder that the piece is provocation as much as prediction.)

The escalation dynamic explained: Once observers adapt to a signal, the signal loses value, and actors must raise the cost to restore credibility. The only way off the escalation treadmill is to establish signals that are both hard to fake and stable enough to resist inflation over time. Neither condition is easily met.

A Taxonomy of Proof-of-Care Methods

The table below organizes the methods Filipp discusses. It spans techniques he presents as already practiced (handwriting, tattoos) and the escalating hypotheticals he uses to make his argument (guild scarification, body subtraction); the latter should be read as speculative rather than documented.

Method	Cost to Producer	Scalability	Spoofability	Durability as Signal	Status in the Essay
Handwriting (digital photo)	Low–Medium	Low	Medium (plotters)	Declining	Presented as practiced today: posts photographed and shared as images
Handwritten flyers (physical)	Medium–High	Very Low	High (pen plotters + gig distribution)	Declining	Presented as practiced today; plotter spoofing is speculative
Tattoos	High (pain + permanence)	Very Low	Very Low	High (personal); Low if mass-adopted	Presented as practiced today; "Grind Hard" and "Meat is murder" examples
Town crying / public speech	Medium (time, vulnerability)	Low	Low–Medium	Medium	Speculative revival imagined in the essay; not an observed trend
Travelling storytellers	Very High (skill + travel)	Very Low	Medium (can be hired)	Medium; guild verification extends it	Speculative; city-to-city, audience-specific framings
Guild scarification	Very High (pain + permanent)	Very Low	Very Low	High (hard to forge at scale)	Speculative worldbuilding: razor cuts + ash; bone-chisel carving
Body subtraction	Extreme (irreversible)	Minimal	Essentially None	Short (escalation pressure continues)	Speculative dystopia: ear/finger severance in fictional warzones

What the Proof-of-Care Crisis Reveals About AI and Human Communication

The methods catalogued above span the grotesque and the mundane, but they share a common architecture: they all work by making the signal costly to produce in a way that AI cannot replicate. The cost is physical, temporal, or social — it lives outside the token space that language models operate in. This is the key insight, and it holds whether one takes Filipp's scenarios literally or as parable: AI cannot feel pain, cannot be permanently marked, cannot be recognized on a street corner. The proof-of-care methods most likely to survive the AI age are those anchored to embodied, irreversible human experience.

This has practical implications for digital communication design. The current default — text, images, video — is now almost entirely within AI's production envelope. If proof of care requires stepping outside that envelope, then the future of trusted communication may be increasingly hybrid: digital distribution of content that carries verifiable analog provenance. Think of a handwritten letter photographed and sent via encrypted message, or a public speech that is livestreamed but can be physically attended and verified in person.

There is also a deeper cultural question here. The escalation toward physical proof-of-care that Filipp dramatizes is, in part, a symptom of institutional failure. When people trust neither platforms nor certificates nor reputations to verify authenticity, they fall back on the oldest verification mechanism available: witnessing a costly, embodied act with their own eyes. This is not a new human impulse — it is as old as oaths sworn on sacred objects, or contracts sealed with blood. AI has not invented this dynamic; it has intensified it by eroding the intermediate layer of textual trust that modernity had built up over centuries.

The challenge for technologists is whether a digital equivalent of costly signaling can be constructed — some form of verifiable human provenance that is as legible and hard to fake as Filipp's imagined guild scar, but does not require anyone to actually be cut. Emerging standards such as the Coalition for Content Provenance and Authenticity (C2PA) and its associated Content Credentials specification attempt part of this: cryptographically signing media at the point of capture, before it touches editing software, so that the chain of custody is machine-verifiable. It is worth noting that such provenance systems prove how a piece of media was captured and edited, not that a human genuinely cared about its content — they attest to origin, not to sincerity. Hardware-level signing built into cameras, invisible watermarking, and trusted-execution environments for content creation are all active research and standardization areas. But none has yet achieved the social legibility of a handwritten note, let alone the visceral credibility of a permanent mark. This remains one of the genuinely open problems of the AI age, and one that builders of human-centered technology will need to reckon with seriously.

---

Key Takeaways

The following points distill the core arguments about proof of care in the age of AI for practitioners, researchers, and communicators who need to act on them now.

• AI has broken the effort signal: The longstanding equation between laborious output and genuine investment has been severed. Any text that could have taken hours to write can now be generated in seconds, making effort invisible as a trust signal.
• Proof of care is the new authenticity problem: Filipp argues the central challenge of communication in the AI age is not quality but verifiable human investment — and that people will reach for costly, hard-to-fake signals to restore it.
• Every signal faces a spoofing attack: In Filipp's framing, pen plotters undermine handwritten flyers and hired storytellers undermine oral authenticity. Any proof-of-care mechanism that scales attracts sophisticated imitation.
• Embodied cost is the hard floor: The methods hardest to fake are those rooted in irreversible physical acts — pain, permanence, and presence. AI cannot replicate these, at least not yet.
• Escalation is a systemic risk: Filipp's darkest thread is that, without stable institutional mechanisms, informal pressure could push toward increasingly extreme demonstrations of sincerity, with no natural ceiling. His most extreme examples are deliberate dystopian illustrations, not documented events.
• Digital provenance tools exist but lack social legibility: Standards such as C2PA and Content Credentials are technically promising but attest to a file's origin rather than a human's sincerity, and have not yet achieved the intuitive, at-a-glance trust that analog proof-of-care signals provide.
• The essay is speculative, and should be read that way: Filipp's town criers, guilds, scarification, and warzone body-subtraction are near-future worldbuilding (with invented place names), not reported trends — a device for stress-testing the argument.
• Historical precedent runs deep: From Caesar's autobiographical dispatches to Homeric bards, the use of embodied, attributed narrative as a trust mechanism is ancient. AI has not created this dynamic; it has made it urgent again.

What Comes Next

The Accelerating Arms Race

The proof-of-care problem will sharpen as generative AI becomes faster, cheaper, and more multimodal — capable of producing not just text but convincing audio, video, and synthetic handwriting at scale. The pen plotter that spoofs a flyer in Filipp's scenario is a crude ancestor of systems that could soon generate convincing "handwritten" letters hard to distinguish from the real thing. Real-time voice cloning already makes phone calls an unreliable authenticity signal. Photorealistic video synthesis is eroding the last refuge of visual evidence. Each capability removed from the "hard to fake" column narrows the space in which analog proof of care can operate and intensifies pressure on the remaining options.

Toward Stable Solutions

This trajectory puts pressure on both technology and culture to develop stable, scalable, and socially legible authenticity signals before the escalation dynamic Filipp describes could drive informal proof-of-care practices toward genuinely dangerous extremes. The most credible path forward involves three interlocking strategies: hardware-anchored content provenance (signing media at the sensor level, before it touches software, as C2PA aims to standardize); strong social norms around disclosure (the cultural expectation that AI-assisted content is labeled, much as health warnings normalized disclosure on other products); and a deliberate cultural revaluation of slowness, presence, and physical commitment as markers of credibility. None of these solutions is sufficient alone. The window for establishing them — before the spoofing arms race runs further ahead — is narrowing. Building the infrastructure of human trust in an AI-saturated world is not a product problem or a policy problem in isolation; it is a civilizational design challenge that will define the next generation of communication.