No. 172 / 339
What is a commercial photographer actually being paid for once a client can generate "a businessman shaking hands" for free?
The shift
Producing a plausible, license-free image of a generic staged scene — the handshake, the team at the whiteboard, the smiling call-center agent — goes from scarce (a commissioned shoot or a stock license) to abundant and near-free. What stays scarce is an image of a specific real subject, the legal releases that only a real shoot produces, and a credible way to prove a given image records something that actually happened.
The axioms
- Producing a technically clean, on-brief image takes trained skill. Rests on image-production craft — lighting, composition, retouch — being scarce.
- A competent version of a generic staged concept has commercial value someone will pay to license. Rests on any usable image of that concept being scarce enough to meter and sell.
- A shoot delivers an image of a specific real thing — this person, this product, this room, this moment. Rests on physical access to a real subject being scarce and non-substitutable.
- A commissioned image can be used commercially because releases tie it to real, consenting subjects and property. Rests on legal model/property releases being obtainable only from a real shoot with real people.
- A named photographer is accountable for the shoot — they showed up, directed it, and stand behind what it depicts. Rests on a person answerable for the capture being scarce.
- A photograph is assumed to record something that was in front of the lens. Rests on the default that making a realistic image required a real scene.
- Stock and commodity shooting is a viable business floor. Rests on volume of competent imagery being scarce enough to price per-asset.
Invalid axioms
- Producing a technically clean, on-brief image takes trained skill. For a generic staged scene, a clean, correctly-lit, well-composed image now exists in seconds for free. The craft that gated who could make a passable commercial shot of nothing in particular is commoditized. Habit-trap: rate cards, portfolios, and pitches still lead with "I can execute to a professional standard," which is now the floor, not the differentiator — for invented subjects it's not even required.
- A competent version of a generic staged concept has commercial value someone will pay to license. The concept handshake, the diverse team high-fiving, the doctor with a tablet — the entire made-up-scene category — no longer needs a shoot or a license, because the buyer can generate it. Habit-trap: shooters still build shot lists, keyword their catalogs, and chase the "corporate generic" brief as bread-and-butter work, metering a thing whose price just went to zero.
- Stock and commodity shooting is a viable business floor. The floor was volume: enough competent, generic frames that a buyer would pay rather than reshoot. That volume is now infinite and free. Habit-trap: photographers treat stock as the safety net under commissioned work and price commissioned work against it, when the net has dropped out. This is the fastest-moving call in the file — the commodity-stock tier is collapsing now, not eventually.
Unchanged axioms
- A shoot delivers an image of a specific real thing. A model can invent a plausible executive; it cannot photograph your CEO, your product on the line, your restaurant's actual plating, or the event that happened on the day. Physical access to a named, real subject is not a prompt — it's a person in a room, a product in hand, a place you were allowed into. The work fleeing into documentary, food, workplace, and product niches is fleeing toward this.
- A commissioned image can be used commercially because releases tie it to real, consenting subjects and property. A signed model or property release is a legal instrument attached to a real person or thing that agreed to appear. Generated imagery has no one to release — which is exactly why it's clean for invented scenes and radioactive for anything a viewer would read as a real, identifiable individual. The release is a scarce artifact only a real shoot produces, and it's becoming part of what the client is actually buying.
- A named photographer is accountable for the shoot. Someone showed up, directed real people, and can attest that the image depicts what it claims. That accountability — answerable if the depiction is challenged, if consent is questioned, if the client needs to prove the shoot occurred — cannot be generated. A model can't be the party who stands behind the capture.
- A photograph is assumed to record something that was in front of the lens. This holds as a thing clients still want — but only as a want, not a default anymore. The value of a real capture is rising precisely because realism no longer implies reality. Calibrate carefully: the belief that any given realistic image is real is already dead (see NEW); what survives is the demand for images that genuinely are, sourced from someone accountable for it.
New axioms
- Proving an image is real is now a separate job from making it. When a generated frame is indistinguishable from a shot one, "this actually happened, with these real people" stops being carried by the image itself and has to be established some other way. The scarce act moves from capture to credible attestation of capture — and no settled mechanism for it exists in commercial practice.
- Provenance and chain-of-custody become part of the deliverable. Clients in regulated, editorial, or trust-sensitive categories increasingly need to show an image is authentic — capture metadata, signed provenance, a documented shoot. This is a capability photographers don't currently sell and mostly can't yet prove, arriving faster than the tooling and norms to support it.
- The commodity tier's collapse strands the people who lived on it. The generic-corporate and stock floor was where new and mid-tier shooters subsidized the access-based work. Removing it doesn't just delete a product line — it removes the on-ramp, and it's unclear what funds the years of shooting that build the access, relationships, and real-subject reputation that stay scarce.
- Distinguishing a real shoot from a convincing fake in the client's own supply chain. Buyers now have to police whether the "photography" they commission or receive is actually photographed — including work passed off by vendors as real. Verifying authorship and authenticity at the point of purchase is a problem clients didn't have when a realistic image implied a camera.
Where it breaks
Shooters and agencies are abandoning the commodity-stock tier because it's worthless now (INVALID: generic staged imagery stopped being scarce), at the same moment that tier was the on-ramp funding the years of work that build real-subject access and trusted authorship — the things that still hold (NEW: the collapse strands the people who'd grow into the scarce work). The floor that trained the next generation of access-based shooters is being removed faster than anything replaces it.
Second collision: the trade still sells "a real photograph" on the old assumption that a realistic image is self-evidently real (INVALID: realism no longer implies reality), while the actual scarce good has become provable reality — releases, accountable authorship, provenance (NEW: proving an image is real is a separate job). Photographers positioned on "I shoot real things" are selling the right asset with none of the machinery to prove it, and clients can't yet tell their genuine work apart from a good fake.
Related axioms
Media
What shifts for creative work with AI?
Media
What changes for film with AI?
Media
What changes for journalism with AI?
Media
What changes for music with AI?
Media
What changes for writing with AI?
Media
When AI can master a track to "radio-ready" in one click, what stays scarce about the audio engineer's ear?
Other axioms
Engineering
Do we still need dedicated data engineers when AI agents can build and self-heal ETL pipelines?
Engineering
What changes for hardware engineering with AI?
Industries
What changes for plumbers with AI?
Engineering
What's left of the "QA engineer" role when developers own their own AI-generated tests?
Product Design
Is pushing pixels still a job when AI generates production-ready screens from a prompt?
Healthcare
Is the pharmacist's checking role obsolete when AI verifies interactions and dispensing, or does accountability keep them?