No, but the mechanical layer of editing has largely gone, and that has changed what editors are paid for. Automatic transcription, silence removal, rough assembly, colour matching and subtitle generation all happen without human involvement now. Federal figures put film and video editors and camera operators at 3 percent growth from 2025 to 2035, about as fast as average, taking the category from 72,000 positions to roughly 73,900 with median pay of $75,100.
Modest growth in a field this exposed to automation is worth explaining, because the reason is not that the tools failed. They work. The reason is that editing was never mainly about the operations.
Key points
- Assembly is automated; storytelling is not. Cutting out silences and syncing audio is solved. Deciding what a sequence should feel like is not.
- Employment grows 3 percent to 2035, adding roughly 1,900 positions from a base of 72,000.
- Volume of video keeps rising, which offsets the efficiency gains within each project.
- The market split. Commodity editing competes directly with software; narrative, commercial and documentary work does not.
- Median pay is $75,100, with a very wide spread that tracks the kind of work rather than years of experience.
What video editors actually do
The job is usually pictured as trimming clips and joining them together. That is the operation, not the work.
Logging and selection. Watching everything shot and deciding what is usable. On a documentary this can be a hundred hours of material for a ninety-minute film, and the selection determines what the finished piece can be.
Structure. Deciding the order, what the audience learns and when, where to withhold information and where to reveal it. This is authorship rather than assembly.
Pacing and rhythm. Judging how long a shot holds, when to cut, how a scene breathes. Two editors given identical material produce recognisably different films, and the difference is almost entirely here.
Performance selection. Choosing between takes based on which reading of a line lands, which frequently means preferring a technically worse take because the performance is better.
Sound and music. How audio carries across cuts, where music enters, how sound shapes emotional response. Often the largest single contributor to how a piece feels.
Client and director collaboration. Interpreting vague notes, proposing alternatives, defending choices, and knowing when a note reflects a real problem in a different place than the one indicated.
The part that did not change
It is worth being precise about what survived, because it defines what an editor should be good at now.
Someone still has to decide what the piece is about. A model asked to cut a sequence will cut a sequence. It has no view on whether this interview subject is more credible than that one, whether the argument holds together, or whether the film is actually about what the brief said it was about. Documentary editors routinely discover the real story in the edit, and that discovery is the job.
Someone still has to judge how something feels. Pacing is not a rule. Holding a shot two seconds longer than comfortable is sometimes exactly right and sometimes a mistake, and the difference depends on everything around it.
Someone still has to handle the client. A note saying the middle is boring may mean the middle is boring, or that the setup failed, or that the client saw a different cut yesterday and is remembering that one. Diagnosing which is a professional skill with no technical component.
Why the craft resists automation more than it appears
There is a specific reason editing holds up better than the tooling suggests, and it is worth naming because it applies to other creative work too.
An edit has no correct answer. A spreadsheet either balances or it does not. A network config either routes traffic or it does not. A cut is a set of choices where several versions are defensible and one is better for reasons that depend on audience, context and intent. Models are trained to produce likely output, and likely is precisely the wrong target when the goal is a piece that stands out from everything else in the feed.
That also explains where automation succeeded fastest. Silence removal has a correct answer. Colour matching between two cameras has a correct answer. Syncing audio to picture has a correct answer. Every solved task in editing is one with an objectively right result, and every unsolved one requires taste.
Where automation genuinely performs
Being honest about this matters more than defending the craft.
Transcription and text-based editing let an editor cut dialogue by deleting words in a transcript, which is dramatically faster than scrubbing a timeline. Silence and filler removal cleans up a talking-head recording in seconds. Automatic reframing converts a horizontal edit to vertical formats. Colour matching balances shots from different cameras. Subtitle generation and translation produce accessible versions without manual timing. Music selection tools suggest tracks that fit a cut’s rhythm. Object removal and rotoscoping handle work that used to take a specialist days.
For simple formats these tools do not assist an editor; they remove the need for one. A podcast clip, a talking-head social video or a straightforward screen recording can now be edited adequately by the person who recorded it, and a large volume of low-complexity work has moved that way.
Why the profession did not shrink
Three factors kept the numbers stable despite that.
Total video volume exploded. Organisations that produced two videos a year now produce two a week. Cheaper editing created demand rather than only satisfying it, and some of that demand still requires someone competent.
Complexity concentrated. As simple work left the professional market, what remained was disproportionately the difficult end: narrative, documentary, commercial and anything where the result has to be good rather than adequate.
Formats multiplied. The same content now ships in several aspect ratios, lengths and platform-specific versions. Automation helps, but somebody has to decide what survives a cut down to fifteen seconds, and that is an editorial judgement each time.
Quality expectations rose with volume. When everyone can produce video, adequate video stops being impressive. Organisations competing for attention find that the bar moved, and clearing it needs the judgement parts rather than the operational ones.
What to know before deciding
| Measure | Film and video editors and camera operators, 2025 |
|---|---|
| Median annual pay | $75,100 |
| Number of jobs | 72,000 |
| Projected growth, 2025 to 2035 | 3 percent (As fast as average) |
| Projected employment change | 1,900 |
| Typical entry-level education | Bachelor’s degree |
Two things about that table matter for a career decision.
The occupation is small at 72,000 positions, and a large share of the work is freelance or project-based. That changes the risk profile considerably compared with salaried employment, and national employment figures understate how variable individual income is.
The wide spread around the median is the important number and it is not in the table. Commodity editing rates have been under sustained pressure, because the alternative now costs a client nothing beyond their own time with free software. Narrative, commercial and high-end documentary rates have held. Which end you work at matters more than the profession-wide average.
There is a related effect on how work is priced. When assembly took two days, an editor could charge for two days. When it takes twenty minutes, charging by time punishes exactly the skill that got faster. Editors who moved to pricing by project or by outcome have generally protected their income better than those still billing hours, because the hours available to bill for have genuinely shrunk while the value delivered has not.
For context on how exposure is measured, the Bureau publishes AI exposure categories for 831 occupations and states plainly that exposure “does not imply job loss, productivity gains, automation probability, or wage effects.” Editing scores as highly exposed, and the flat-to-modest growth is a useful reminder of what that measure leaves out.
What actually changes over the next decade
- Assembly time collapses. Rough cuts arrive in minutes, which shifts value toward structure and judgement.
- Simple work leaves the professional market. Clients edit their own straightforward content, and that is not coming back.
- Format versioning becomes routine. One edit ships in many shapes, largely automatically.
- Junior entry gets harder. Assembly and logging were how editors learned, and both are the parts most automated.
- Sound work holds value. It is less automatable than picture and contributes disproportionately to quality.
- Client expectations compress. Faster turnaround is assumed, which pressures the time available for the judgement parts.
Decision framework
Five questions if you edit video now.
- What share of your work is assembly? Logging, syncing, silence removal and rough cutting are the exposed categories, and they are also the easiest to measure.
- Which end of the market do you work at? Commodity content competes with free tools. Work where quality is the point does not.
- Can you defend a structural choice? Being able to explain why a sequence works in this order is what makes an editor an author rather than an operator.
- Do you handle clients well? Interpreting vague notes and proposing alternatives is a large part of what clients actually pay for.
- Will you use the tools deliberately? Editors who automate assembly and spend the recovered hours on structure produce better work than either those who refuse the tools or those who deliver the first automatic cut.
That last habit transfers beyond editing. Knowing what a tool does well, where its confident output is wrong, and how to spend the time it saves, applies anywhere these systems are used. If you want a structured route in, explore Coursiv AI lessons and check current plan details on the official site.
Your next step
If you edit now, time one project honestly and split it between assembly, structural decisions, and client handling. That ratio is the clearest measure of your own exposure, and most editors are surprised how much of the schedule went to operations that are now near-instant.
Then take a project where you own the structure rather than executing someone else’s. The ability to point at a finished piece and explain why it is shaped that way is what moves an editor from the priced-against-software end of the market to the other one.