Are AI-generated images legal to use for business in 2026

The US Supreme Court settled part of this debate in March 2026. The rest? Still messy — and getting messier.

When the Supreme Court declined to hear Thaler v. Perlmutter, it locked in a clear position: pure AI-generated content can’t be copyrighted under US law. No copyright owner. Not you, not Midjourney, not OpenAI. That ruling rippled through every legal team, creative agency, and SaaS startup using AI imagery for commercial work. But “you can’t copyright it” doesn’t mean “you can freely use it for business.” Those are two different questions, and conflating them is where most companies get into trouble.

So — are AI-generated images legal to use for business in 2026? The short answer is yes, with significant conditions. The longer answer requires understanding three separate layers: copyright law, platform licensing terms, and trademark/similarity risk. Miss any layer and you’re exposed.

Key Takeaways

• The March 2026 Supreme Court decision in Thaler v. Perlmutter confirmed that pure AI-generated images receive no copyright protection under US law, placing them effectively in the public domain.
• Commercial rights for AI-generated images are contractual, not copyright-based — meaning platform terms of service govern what you can actually sell or publish.
• Adobe Firefly offers the strongest enterprise protection in 2026, including indemnification; Midjourney requires a $30+/month paid plan for any commercial use.
• Individual user liability risk rises significantly when prompting AI to reproduce recognizable copyrighted characters, styles, or brand elements for commercial output.
• Over 70 active AI copyright lawsuits as of mid-2026 target AI companies directly, not end users — but that doesn’t make end users risk-free.

The Legal Landscape: What the March 2026 Ruling Actually Changed

The Thaler v. Perlmutter case had been grinding through the US court system for years. Stephen Thaler argued that his AI system, DABUS, should be recognized as the copyright author of works it generated autonomously. The Supreme Court’s March 2026 refusal to hear the case ended that argument definitively at the federal level.

The US Copyright Office’s position — unchanged and now further reinforced — is that copyright requires meaningful human creative decision-making. A prompt isn’t enough. Even a detailed, 400-word prompt with style references, negative weights, and aspect ratio specifications doesn’t cross that threshold. The Office treats prompts as instructions, not creative authorship.

What can be protected? The human-authored portions of a hybrid workflow. If a designer generates a base image with Midjourney, then spends two hours in Photoshop making iterative selections, painting over sections, adding original illustrated elements, and documenting those creative decisions — that documented human layer might qualify for protection. The AI-generated substrate won’t. This distinction matters enormously for businesses building visual brand assets in 2026.

The copyright gap also creates a competitive exposure problem most teams haven’t fully priced in. Content that can’t be copyrighted can be freely copied. Any competitor can take your AI-generated hero image and use it too, legally. That’s not a hypothetical edge case. That’s the default state of any unmodified AI output you publish commercially.

Platform Licensing: Where the Real Rules Live

Copyright law tells you who owns something. Platform terms tell you what you’re allowed to do with it. For businesses, the platform layer is often more immediately relevant — and more frequently ignored.

According to the Red Escuela AI Copyright Guide, commercial rights to AI-generated images are contractual in nature. They flow from the platform’s terms of service, not from any copyright claim you hold. Each major tool handles this differently.

Comparison: Commercial Licensing Across Major AI Image Platforms (2026)

Adobe Firefly stands apart from the field. It was built specifically on licensed datasets — Adobe Stock imagery with proper rights clearances — and backs enterprise use with indemnification. That means if a third party sues your business over a Firefly-generated image, Adobe covers your legal defense. No other major platform offers that in 2026.

Midjourney’s revenue threshold clause trips up more companies than any other term in this space. Free and Basic tier outputs technically prohibit commercial use. And if your business crossed $1M in annual revenue, the Standard plan isn’t sufficient — you need Pro ($60/month) or Mega ($120/month). Most startups don’t notice this until they’re already in violation. That’s not a fine-print technicality. That’s a material licensing breach.

Liability Zones: Where Individual Risk Actually Lives

The 70+ active AI copyright lawsuits as of mid-2026 mostly target AI companies, not their users. Getty Images vs. Stability AI, The New York Times vs. OpenAI, and similar cases are upstream disputes about training data — not downstream disputes about what end users generated.

But “the lawsuits aren’t aimed at users” doesn’t mean user risk is zero. It means the exposure is different in kind, not absent.

The highest-risk scenario for any business: prompting an AI to generate a recognizable copyrighted character — a clearly identifiable Disney figure, a Marvel hero, a branded mascot — and selling that output commercially. Substantial similarity standards apply regardless of whether AI was involved in creation. A human drawing and selling unlicensed Mickey Mouse art faces infringement risk. AI-generated output that looks substantially like the same character faces the same exposure. The tool doesn’t change the analysis.

Trademark risk is a second, underappreciated lane. Generating imagery that looks confusingly similar to a competitor’s logo or trade dress — even accidentally — creates trademark exposure that copyright law doesn’t touch. This approach can fail badly when teams use competitor brands as reference points in their prompts without thinking through what the output might actually resemble.

What Businesses Should Actually Do Right Now

Scenario 1: Startup using AI imagery for marketing collateral The risk is low if you’re using Firefly or DALL-E and staying away from recognizable IP. Audit your prompts for character references or brand-adjacent imagery. Document your creative process if you’re layering human edits — that documentation is your best protection if questions arise later. This isn’t always the answer for every use case, but for standard marketing work, it’s close.

Scenario 2: Agency producing AI images for client deliverables Client contracts need explicit AI disclosure clauses now. Who owns the output? Who’s liable if a third party raises a similarity claim? These aren’t hypothetical — they’re billable conversations you should be having before the project starts, not after the invoice. Adobe Firefly’s indemnification makes it the default recommendation for client-facing commercial work.

Scenario 3: Enterprise team scaling AI visual production Revenue thresholds on Midjourney mean your licensing cost structure changes as you grow. Run a platform audit now. If you’re approaching $1M ARR and still on Standard tier Midjourney, you may already be operating outside the terms. Switch tiers or switch platforms — both are straightforward. Discovering the violation during due diligence for a funding round is not.

What to watch in Q4 2026:

• Whether EU AI Act implementation creates disclosure requirements affecting how AI-generated images must be labeled in commercial contexts
• Any Copyright Office rulemaking responding to the Thaler outcome — formal guidance on hybrid human-AI workflows is still pending and could expand protection for documented collaborative processes

Where This Is Headed

The legal picture around AI-generated images for business has clarified substantially in 2026, but it hasn’t simplified.

The Supreme Court’s Thaler refusal settled the copyright ownership question: AI-generated content is effectively unprotectable without meaningful human creative input. Commercial rights flow from platform contracts, not copyright claims — and those contracts vary significantly. Adobe Firefly remains the safest enterprise choice. Midjourney’s revenue thresholds catch growing companies off guard more than any other compliance gap in this space. User liability risk concentrates around identifiable IP reproduction and trademark similarity, not general AI image use.

Over the next 6-12 months, expect EU AI Act labeling requirements to add compliance overhead for businesses operating in European markets. Expect the Copyright Office to issue more specific guidance on hybrid workflows. And expect platform terms to get more granular as commercial stakes rise and litigation outcomes accumulate.

The bottom line: AI-generated images are legal to use for business in 2026 — if you pick the right platform, stay on the right plan tier, and keep your prompts away from recognizable third-party IP. That’s not a complex framework. It just requires actually reading the terms before you publish, not after something goes wrong.

What platform is your team currently using for commercial AI image work — and have you audited whether your current plan tier matches your revenue level?

References

1. Can You Use AI and ChatGPT Images Commercially in 2026
2. AI and Copyright: What You Can and Can’t Do with AI-Generated Content (2026 Guide) | Red Escuela
3. Can AI-Generated Images & Videos Be Used Commercially? | getimg.ai

Photo by Franck V. on Unsplash