Meta's new AI image tool can use public Instagram photos to generate images for other users, The Hacker News reports. The platform where much of the working photography industry keeps its public portfolio has, in effect, opened that portfolio to a generator anyone can prompt.

The report establishes three facts. The tool exists. It generates images. The pool of photos it draws on includes public Instagram posts — other people's, not just the prompter's own. What the report leaves out matters just as much for anyone running a photo business on the platform. No launch date. No list of markets. No opt-out mechanism. No royalty or fee structure. No clarification on whether public photos feed model training or enter the pipeline only when a user generates an image.

The revenue exposure is direct. For most working photographers, a public Instagram feed is a client-acquisition channel first and a portfolio second. Under the reported arrangement, every image posted to attract bookings also functions as material for a generator that competitors, art buyers, and existing clients can prompt. Wedding, editorial, and commercial shooters who post selects publicly face the sharpest version of the problem: the images that win them jobs are the ones the tool can use. The commercial risk is simple — if the tool can build images from your public photos, it can produce images that resemble your work without you in the transaction. That is a substitution risk to the assignment and licensing revenue the feed exists to create.

The operative boundary is public. The report's framing centers on public photos, which implies private accounts and private posts sit outside the tool's reach. The report does not confirm that, and the distinction is the first question to put to Meta before changing any account settings.

The workflow decisions that follow are familiar from training-data disputes, with one difference: this use is generative and on demand, not statistical. If a photo can enter someone else's image at prompt time, the exposure runs continuously for as long as the post stays public. The options narrow to three. Lock the account and lose the discovery channel. Curate what goes public and accept partial exposure. Keep posting and price the risk into rates. Each choice carries a client-acquisition cost, and none comes with compensation attached — the report names no royalty, license, or payment of any kind for the photographers whose work supplies the tool.

Treat the announcement as data, not settled policy. The report describes a tool; it does not show measured results, and it does not say whether the feature is live for all users, a subset, or a test cohort. Announcements of this kind have preceded both full rollouts and quiet retreats. Until Meta publishes specifications — which photos qualify, how generation draws on them, what provenance the outputs carry, whether commercial use of those outputs faces restriction — photographers must price an unquantified risk.

The questions worth tracking are concrete. Does an opt-out exist, and does using it cost reach? Do images generated from a photographer's photos carry labels or provenance data? Can a public archive be excluded wholesale? The report answers none of these. The next move belongs to Meta: publish the tool's terms, or leave the industry pricing the ambiguity into how it uses the public side of the platform.